Ideas, experiments & evidence.

Abstract

Large language models (LLMs) can generate fluent Arabic answers, yet factual errors remain difficult to detect, localize, explain, and verify. Existing hallucination benchmarks often provide response-level labels, with limited support for identifying the exact erroneous content, explaining why it is incorrect, or selecting the correct factual answer. We introduce HalluTruthQA, a fine-grained benchmark for hallucination evaluation in Arabic question answering. The benchmark contains 2,400 expert-curated examples across four knowledge-intensive domains: Islamic knowledge, history, science, and geography. Each example pairs an Arabic question and a model-generated answer with a verified reference answer, a binary hallucination label, six candidate answers for factual verification, and, for hallucinated answers, character-level erroneous spans, human-written explanations, and macro and micro hallucination types. We evaluate four open-source LLMs, Allam, Falcon-H1, Qwen32, and Silma, in a zero-shot setting across hallucination detection, span-level localization, factual verification, and explanation evaluation. Results show that these tasks capture different abilities: no single model achieves the strongest performance across all tasks, with best scores of 0.880 Macro-F1 for detection, 0.516 F1-Sp for localization, 0.852 LO-Score for factual verification, and 0.644 final score for explanation evaluation. Our taxonomy shows that hallucination evaluation should move beyond detection toward localizing, verifying, and explaining factual errors. The code, dataset, prompts, and evaluation scripts are available at https://gitlab.com/nlpresearcher/HalluTruthQA.

BibTeX

@misc{Bouchekif2026HalluTruthQA,
  title         = {HalluTruthQA: A Fine-Grained Benchmark for Hallucination Detection, Localization, and Explanation in Arabic Question Answering},
  author        = {Abdessalam Bouchekif and Mohammed-En-Nadhir Zighem and Salah Eddine Bekhouche and Hichem Telli and Somaya Eltanbouly and Shahd Gaben and Heba Sbahi and Samer Rashwani and Mutaz Al-Khatib and Emad Mohamed and Mohammed Ghaly and Abdenour Hadid},
  journal       = {arXiv preprint arXiv:2607.20219},
  year          = {2026},
  eprint        = {2607.20219},
  archivePrefix = {arXiv},
  url           = {https://arxiv.org/abs/2607.20219}
}

Abstract

Ambivalence and hesitancy (A/H) are subtle affective states where a person shows conflicting signals through different channels -- saying one thing while their face or voice tells another story. Recognising these states automatically is valuable in clinical settings, but it is hard for machines because the key evidence lives in the disagreements between what is said, how it sounds, and what the face shows. We present ConflictAwareAH, a multimodal framework built for this problem. Three pre-trained encoders extract video, audio, and text representations. Pairwise conflict features -- element-wise absolute differences between modality embeddings -- serve as bidirectional cues: large cross-modal differences flag A/H, while small differences confirm behavioural consistency and anchor the negative class. This conflict-aware design addresses a key limitation of text-dominant approaches, which tend to over-detect A/H (high F1-AH) while struggling to confirm its absence: our multimodal model improves F1-NoAH by +4.6 points over text alone and halves the class-performance gap. A complementary text-guided late fusion strategy blends a text-only auxiliary head with the full model at inference, adding +4.1 Macro F1. On the BAH dataset from the ABAW10 Ambivalence/Hesitancy Challenge, our method reaches 0.694 Macro F1 on the labelled test split and 0.715 on the private leaderboard, outperforming published multimodal baselines by over 10 points -- all on a single GPU in under 25 minutes of training.

BibTeX

@misc{Bekhouche2026ConflictAware,
  title  = {Conflict-Aware Multimodal Fusion for Ambivalence and Hesitancy Recognition},
  author = {Salah Eddine Bekhouche and Hichem Telli and Azeddine Benlamoudi and Salah Eddine Herrouz and Abdelmalik Taleb-Ahmed and Abdenour Hadid},
  journal= {arXiv preprint arXiv:2603.15818},
  year   = {2026},
  eprint = {2603.15818},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:P5F9QuxV20EC}
}

Abstract

Driver fatigue significantly contributes to road accidents and fatalities, particularly during late-night or early-morning travel due to lack of sleep. Fatigue impairs attention and delays reaction times, increasing accident risk. The American Automobile Association estimates that drowsy driving leads to over 320,000 accidents annually, causing approximately 6,400 fatal collisions. Building effective drowsiness detection systems has therefore become a paramount need. This paper introduces a novel neural solution for drowsiness detection from the driver’s face using a semi-supervised learning technique. The framework integrates YOLOv8 for accurate face detection and the Swin Transformer to extract subtle fatigue cues using hierarchical attention. An adaptive pseudo-labeling strategy with dynamic confidence thresholding reliably expands the training set, directly improving model robustness. Comprehensive …

BibTeX

@article{Rahmani2026ASemi,
  title  = {A Semi-Supervised Neural Framework for Real-Time Drowsiness Detection Using Facial Cues},
  author = {Chahrazad Rahmani and Azeddine Benlamoudi and Yazid Bounab and Salah Eddine Bekhouche and Djamel Samai and Fadi Dornaika and Abdel Taleb and Samir Brahim Belhaouari},
  journal= {IEEE Access},
  publisher= {IEEE},
  year   = {2026},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:f2IySw72cVMC}
}

Abstract

Medical image segmentation is crucial for many healthcare tasks, including disease diagnosis and treatment planning. One key area is the segmentation of skin lesions, which is vital for diagnosing skin cancer and monitoring patients. In this context, this paper introduces SegDT, a new segmentation model based on Diffusion Transformer (DiT). SegDT is designed to work on low-cost hardware and incorporates Rectified Flow, which improves generation quality at reduced inference steps and maintains the flexibility of standard diffusion models. Our method is evaluated on three benchmarking datasets and compared against several existing works, achieving state-of-the-art results while maintaining fast inference speeds. This makes the proposed model appealing for real-world medical applications. This work advances the performance and capabilities of deep learning models in medical image analysis, enabling faster, more accurate diagnostic tools for healthcare professionals.

BibTeX

@inproceedings{bekhouche2026segdt,
  title     = {SegDT: A Diffusion Transformer-Based Segmentation Model for Medical Imaging},
  author    = {Bekhouche, Salah Eddine and Maroun, Gaby and Dornaika, Fadi and Hadid, Abdenour},
  booktitle = {Image Analysis and Processing -- ICIAP 2025},
  series    = {Lecture Notes in Computer Science},
  volume    = {16167},
  pages     = {54--66},
  publisher = {Springer Nature Switzerland},
  year      = {2026},
  doi       = {10.1007/978-3-032-10185-3_5}
}

Abstract

Accurate classification of hyperspectral imagery (HSI) is often frustrated by the tension between high-dimensional spectral data and the extreme scarcity of labeled training samples. While hierarchical models like LoLA-SpecViT have demonstrated the power of local windowed attention and parameter-efficient fine-tuning, the quadratic complexity of standard Transformers remains a barrier to scaling. We introduce VP-Hype, a framework that rethinks HSI classification by unifying the linear-time efficiency of State-Space Models (SSMs) with the relational modeling of Transformers in a novel hybrid architecture. Building on a robust 3D-CNN spectral front-end, VP-Hype replaces conventional attention blocks with a Hybrid Mamba-Transformer backbone to capture long-range dependencies with significantly reduced computational overhead. Furthermore, we address the label-scarcity problem by integrating dual-modal Visual and Textual Prompts that provide context-aware guidance for the feature extraction process. Our experimental evaluation demonstrates that VP-Hype establishes a new state of the art in low-data regimes. Specifically, with a training sample distribution of only 2%, the model achieves Overall Accuracy (OA) of 99.69% on the Salinas dataset and 99.45% on the Longkou dataset. These results suggest that the convergence of hybrid sequence modeling and multi-modal prompting provides a robust path forward for high-performance, sample-efficient remote sensing.

BibTeX

@misc{Sellam2026VPHype,
  title  = {VP-Hype: A Hybrid Mamba-Transformer Framework with Visual-Textual Prompting for Hyperspectral Image Classification},
  author = {Abdellah Zakaria Sellam and Fadi Abdeladhim Zidi and Salah Eddine Bekhouche and Ihssen Houhou and Marouane Tliba and Cosimo Distante and Abdenour Hadid},
  journal= {arXiv preprint arXiv:2603.01174},
  year   = {2026},
  eprint = {2603.01174},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:abG-DnoFyZgC}
}

Abstract

Estimating a person’s age from a single facial image remains challenging because ageing cues are subtle, non-linear, and easily confounded by identity, pose, illumination, and demographic factors. In this work, we investigate the suitability of the Global Context Vision Transformer (GC-ViT), originally introduced for general vision tasks, as a pure transformer backbone for facial age estimation. We extend GC-ViT to scalar age regression through a lightweight MLP head and a task-specific training strategy, with the emphasis placed on a targeted reformulation of the architecture for facial age estimation and on its comprehensive empirical validation in this setting. GC-ViT is well suited to age estimation because it combines short-range window self-attention with long-range global attention, enabling a single hierarchical transformer to capture both fine texture cues (e.g., wrinkles) and global facial proportions. To …

BibTeX

@article{Maroun2026AGlobal,
  title  = {A Global-Context Vision Transformer for Facial Age Estimation},
  author = {G Maroun and SE Bekhouche and J Charafeddine and F Dornaika},
  journal= {Cognitive Computation},
  volume = {18},
  number = {1},
  pages  = {72},
  publisher= {Springer US},
  year   = {2026},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:SP6oXDckpogC}
}

Abstract

We present AffectFlow-DINO, a multi-task learning system for the 11th ABAW challenge that extends a standard deterministic architecture with a conditional rectified-flow head to model the inherent ambiguity of in-the-wild facial behavior. Instead of predicting a single affect estimate, the model learns a conditional generative distribution, enabling uncertainty-aware one-to-many predictions through Monte Carlo sampling. The system jointly estimates continuous valence-arousal, classifies eight facial expressions, and detects twelve Action Units from static face images. Built on a frozen DINOv3 ViT-S/16 backbone, extensive ablation studies show that rectified-flow decoding consistently improves deterministic prediction, particularly for valence-arousal estimation (CCC-V ). We further show that post-hoc threshold calibration effectively recovers performance on severely imbalanced rare classes (e.g., Fear: ) without retraining. Combined with backbone fine-tuning and flow retuning, the final model achieves , substantially outperforming the official challenge baseline of .

BibTeX

@misc{Bekhouche2026AffectFlowDINO,
  title  = {AffectFlow-DINO: Uncertainty-Aware Multi-Task Affect Estimation via Conditional Rectified Flow},
  author = {Salah Eddine Bekhouche and Abdellah Zakaria Sellam and Fadi Dornaika and Abdenour Hadid},
  journal= {arXiv preprint arXiv:2607.13250},
  year   = {2026},
  eprint = {2607.13250},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:OU6Ihb5iCvQC}
}

Abstract

Age estimation based on facial images is a challenging task due to the complex and nonlinear nature of facial aging, which is influenced by both genetic and environmental factors. To address this challenge, we propose a hybrid ConvNeXt–Transformer framework that combines convolutional local feature extraction with attention-based global contextual modeling within a unified age regression pipeline. The methodological contribution of this work lies in the sequential integration of these two complementary paradigms for facial age estimation, allowing the model to capture both fine-grained textural cues—such as wrinkles and skin spots—and long-range spatial dependencies. We evaluate the proposed framework on benchmark datasets including MORPH II, CACD, UTKFace, and AFAD. The results show competitive performance across these datasets and confirm the effectiveness of the proposed hybrid design …

BibTeX

@article{Maroun2026ConvNeXtMeets,
  title  = {ConvNeXt Meets Vision Transformers: A Powerful Hybrid Framework for Facial Age Estimation},
  author = {Gaby Maroun and Salah Eddine Bekhouche and Fadi Dornaika},
  journal= {Applied Sciences},
  volume = {16},
  number = {7},
  pages  = {3281},
  publisher= {MDPI},
  year   = {2026},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:1sJd4Hv_s6UC}
}

Abstract

Islamic inheritance (Ilm al-Mawarith) is a multi-stage legal reasoning task requiring the identification of eligible heirs, resolution of blocking rules (hajb), assignment of fixed and residual shares, handling of adjustments such as awl and radd, and generation of a consistent final distribution. The task is further complicated by variations across legal schools and civil-law codifications, requiring models to operate under explicit legal configurations. We present a retrieval-augmented generation (RAG) pipeline for this setting, combining rule-grounded synthetic data generation, hybrid retrieval (dense and BM25) with cross-encoder reranking, and schema-constrained output validation. A symbolic inheritance calculator is used to generate a large high-quality synthetic corpus with full intermediate reasoning traces, ensuring legal and numerical consistency. The proposed system achieves a MIR-E score of 0.935 and ranks first on the official QIAS 2026 blind-test leaderboard. Results demonstrate that retrieval-grounded, schema-aware generation significantly improves reliability in high-precision Arabic legal reasoning tasks.

BibTeX

@misc{Swaileh2026CVPDat,
  title  = {CVPD at QIAS 2026: RAG-Guided LLM Reasoning for Al-Mawarith Share Computation and Heir Allocation},
  author = {Wassim Swaileh and Mohammed-En-Nadhir Zighem and Hichem Telli and Salah Eddine Bekhouche and Abdellah Zakaria Sellam and Fadi Dornaika and Dimitrios Kotzinos},
  journal= {arXiv preprint arXiv:2603.24012},
  year   = {2026},
  eprint = {2603.24012},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:CHSYGLWDkRkC}
}

Abstract

Accurate medical image segmentation requires both long-range contextual reasoning and precise boundary delineation, a task where existing transformer- and diffusion-based paradigms are frequently bottlenecked by quadratic computational complexity and prohibitive inference latency. We propose RF-HiT, a Rectified Flow Hierarchical Transformer that integrates an Hourglass Transformer backbone with a multi-scale hierarchical encoder for anatomically guided feature conditioning. Unlike prior diffusion-based approaches that rely on hundreds of denoising steps, RF-HiT leverages rectified flow with efficient transformer blocks, achieving linear complexity and requiring only a few discretization steps. The model further fuses conditioning features at each resolution via learnable interpolation, enabling effective multi-scale feature integration with minimal computational overhead. As a result, RF-HiT achieves a strong efficiency-performance trade-off, requiring only 10.14 GFLOPs, 13.6M parameters, and inference in as few as 3 steps. Despite its compact design, RF-HiT attains 91.27% mean Dice on ACDC and 87.40% on BraTS 2021, achieving performance comparable to or exceeding that of significantly more intensive architectures. These results suggest that RF-HiT is a promising, computationally efficient foundation for clinical image segmentation.

BibTeX

@misc{Djouama2026RFHiT,
  title  = {RF-HiT: Rectified Flow Hierarchical Transformer for General Medical Image Segmentation},
  author = {Ahmed Marouane Djouama and Abir Belaala and Abdellah Zakaria Sellam and Salah Eddine Bekhouche and Cosimo Distante and Abdenour Hadid},
  journal= {arXiv preprint arXiv:2604.19570},
  year   = {2026},
  eprint = {2604.19570},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:KxtntwgDAa4C}
}

Abstract

Accurate quantification of lung disease severity from chest imaging is critical for clinical decision-making and resource allocation. We propose a tri-modal deep learning framework, TMF-RSE (Tri-Modal Fusion with Regional Semantics and Evidential Uncertainty), that combines appearance features from two-dimensional chest inputs, structural features from lung segmentation masks, and semantic features from vision-language models (VLMs) for severity quantification. Our approach employs complementary fusion mechanisms that integrate semantic guidance, structural priors, and hierarchical interactions across modalities. The model employs evidential regression to provide both severity predictions and uncertainty estimates. Experiments on the Per-COVID-19 CT and RALO datasets show that TMF-RSE outperforms recent transformer-based baselines, achieving MAE of 4.02 and Pearson correlation of 0.9629 on Per-COVID-19 validation, and 0.339 MAE / 0.973 PC on RALO geographic extent.

BibTeX

@misc{Zidi2026TMFRSE,
  title  = {TMF-RSE: Tri-Modal Fusion with Regional Semantics and Evidential Uncertainty for Lung Severity Scoring},
  author = {Fadi Abdeladhim Zidi and Salah Eddine Bekhouche and Abdellah Zakaria Sellam and Gaby Maroun and Fadi Dornaika and Cosimo Distante},
  journal= {arXiv preprint arXiv:2607.06356},
  year   = {2026},
  eprint = {2607.06356},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:uWQEDVKXjbEC}
}

Abstract

Hyperspectral image classification remains a challenging task due to the high dimensionality of spectral data, significant inter-band redundancy, and the limited availability of annotated samples. While recent transformer-based models have improved the global modeling of spectral–spatial dependencies, their scalability and adaptability under label-scarce conditions remain limited. In this work, we propose LoLA-SpecViT (Low-Rank Adaptation Local Attention Spectral Vision Transformer), a lightweight spectral vision transformer that addresses these limitations through a parameter-efficient architecture tailored to the unique characteristics of hyperspectral imagery. Our model combines a 3D convolutional spectral front-end with local window-based self-attention, enhancing both spectral feature extraction and spatial consistency while reducing computational complexity. To further improve adaptability, we integrate enhanced Low-Rank Adaptation (LoRA) into attention and projection layers, enabling fine-tuning with over 70% fewer trainable parameters. A novel cyclical learning rate scheduler modulates LoRA adaptation strength during training, improving convergence and generalization. Extensive experiments on four benchmark datasets—WHU-Hi LongKou, WHU-Hi HongHu, Salinas, and QUH-Qingyun—demonstrate that LoLA-SpecViT consistently outperforms state-of-the-art baselines, achieving up to 99.91% accuracy with substantially fewer parameters and enhanced robustness under low-label regimes. The proposed framework provides a scalable and generalizable solution for real-world HSI applications in agriculture, environmental monitoring, and remote sensing analytics.

BibTeX

@article{zidi2025lolaspecvit,
  title   = {LoLA-SpecViT: Local Attention SwiGLU Vision Transformer with LoRA for Hyperspectral Imaging},
  author  = {Zidi, Fadi Abdeladhim and Boukhari, Djamel Eddine and Sellam, Abdellah Zakaria and Ouafi, Abdelkrim and Distante, Cosimo and Bekhouche, Salah Eddine and Taleb-Ahmed, Abdelmalik},
  journal = {International Journal of Applied Earth Observation and Geoinformation},
  volume  = {144},
  pages   = {104924},
  year    = {2025},
  doi     = {10.1016/j.jag.2025.104924}
}

Abstract

Age estimation from facial images is a challenging computer vision task with diverse real-world applications. This paper proposes a novel hybrid architecture that integrates ConvNeXt, a state-of-the-art convolutional neural network (CNN), with a Vision Transformer (ViT) to leverage the complementary strengths of localized feature extraction and global attention. The proposed ConvNeXt-ViT hybrid is evaluated on benchmark age estimation datasets (MORPH II, CACD, AFAD, and IMDB-Clean) and achieves superior performance in terms of mean absolute error (MAE) compared to existing approaches. We explore various architectural configurations, use pre-trained models to address data limitations, and apply advanced training strategies and regularization techniques. Comprehensive ablation studies highlight the impact of each component and training strategy, particularly the importance of adapted attention …

BibTeX

@article{Maroun2025IntegratingConvNeXt,
  title  = {Integrating ConvNeXt and vision transformers for enhancing facial age estimation},
  author = {G Maroun and SE Bekhouche and J Charafeddine and F Dornaika},
  journal= {Computer Vision and Image Understanding},
  pages  = {104542},
  publisher= {Academic Press},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:a0OBvERweLwC}
}

Abstract

Islamic inheritance law (’Ilm al-Mawarıth) requires precise identification of heirs and calculation of shares, which poses a challenge for AI. In this paper, we present a lightweight framework for solving multiple-choice inheritance questions using a specialised Arabic text encoder and Attentive Relevance Scoring (ARS). The system ranks answer options according to semantic relevance, and enables fast, on-device inference without generative reasoning. We evaluate Arabic encoders (MARBERT, ArabicBERT, AraBERT) and compare them with API-based LLMs (Gemini, DeepSeek) on the QIAS 2025 dataset. While large models achieve an accuracy of up to 87.6%, they require more resources and are contextdependent. Our MARBERT-based approach achieves 69.87% accuracy, presenting a compelling case for efficiency, on-device deployability, and privacy. While this is lower than the 87.6% achieved by the best-performing LLM, our work quantifies a critical trade-off between the peak performance of large models and the practical advantages of smaller, specialized systems in high-stakes domains.

BibTeX

@inproceedings{Bekhouche2025CVPDat,
  title  = {CVPD at QIAS 2025 Shared Task: An Efficient Encoder-Based Approach for Islamic Inheritance Reasoning},
  author = {Salah Eddine Bekhouche and Abdellah Zakaria Sellam and Telli Hichem and Cosimo Distante and Abdenour Hadid},
  booktitle= {Proceedings of The Third Arabic Natural Language Processing Conference: Shared Tasks},
  pages  = {929-934},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:SeFeTyx0c_EC}
}

Abstract

Arabic poses a particular challenge for natural language processing (NLP) and information retrieval (IR) due to its complex morphology, optional diacritics and the coexistence of Modern Standard Arabic (MSA) and various dialects. Despite the growing global significance of Arabic, it is still underrepresented in NLP research and benchmark resources. In this paper, we present an enhanced Dense Passage Retrieval (DPR) framework developed specifically for Arabic. At the core of our approach is a novel Attentive Relevance Scoring (ARS) that replaces standard interaction mechanisms with an adaptive scoring function that more effectively models the semantic relevance between questions and passages. Our method integrates pre-trained Arabic language models and architectural refinements to improve retrieval performance and significantly increase ranking accuracy when answering Arabic questions. The code …

BibTeX

@inproceedings{Bekhouche2025Enhancedarabic,
  title  = {Enhanced arabic text retrieval with attentive relevance scoring},
  author = {Salah Eddine Bekhouche and Azeddine Benlamoudi and Yazid Bounab and Fadi Dornaika and Abdenour Hadid},
  booktitle= {2025 IEEE 35th International Workshop on Machine Learning for Signal Processing (MLSP)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:70eg2SAEIzsC}
}

Abstract

Multi-Object Tracking (MOT) is a crucial component of autonomous driving, enabling vehicles to perceive and predict dynamic environments accurately. In this work, we propose a multi-object tracking framework that integrates a CNN-based YOLOv11 detector with the DeepSORT tracker under the Tracking By Detection paradigm. To the best of our knowledge, this is the first integration of YOLOv11 and DeepSORT in a 2D Intelligent Transportation Systems (ITS) tracking pipeline. Our method enhances tracking performance by leveraging YOLOv11’s advanced feature extraction for high-precision object detection, while DeepSORT ensures robust identity preservation across frames. Our approach demonstrates superior tracking performance in complex urban driving scenarios, surpassing existing state-of-the-art methods. Specifically, our model achieves a MOTA of 56.61% and an HOTA of 56.28 % on the KITTI …

BibTeX

@inproceedings{Kezzal2025ImprovingAutonomous,
  title  = {Improving Autonomous Driving Perception: A Multi-Object Tracking Framework with CNN and DeepSORT},
  author = {Chahira Kezzal and Selsabil Benderradji and Azeddine Benlamoudi and Salah Eddine Bekhouche},
  booktitle= {2025 7th International Conference on Pattern Analysis and Intelligent Systems (PAIS)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:HoB7MX3m0LUC}
}

Abstract

Deep learning, particularly through sophisticated architectures like CNN, has significantly advanced automated facial analysis. Tasks such as recognition and attribute analysis have seen performance boosts. However, achieving truly robust and versatile systems, especially for complex regression tasks like age estimation or dynamic state assessments like driver drowsiness detection, faces persistent challenges. A key hurdle remains developing models that reliably handle extreme variations in real-world conditions including pose, illumination, expression, occlusions, and intrinsic image quality issues. While techniques like attention mechanisms and specialized network designs exist, accurately interpreting subtle age-related facial changes across a lifetime or detecting fine-grained behavioral cues indicative of drowsiness under these variations remains difficult. Furthermore, ensuring robust generalization across diverse demographics, unseen environments, and varying data acquisition setups often requires more than standard data augmentation or transfer learning, demanding tailored methodological innovations.Addressing these specific challenges in facial age estimation and driver drowsiness detection is crucial. Current age estimation methods, often relying on direct regression or simple classification with standard CNNs, can struggle with the non¬ linear nature of aging, sensitivity to variations unrelated to age, and may not adequately capture distinct features relevant to different life stages. Similarly vision-based drowsiness detection often relies on indicators like PERCLOS, yawn frequency, or head pose, typically extracted using …

BibTeX

@misc{Bekhouche2025AdvancingFace,
  title  = {Advancing Face Analysis in Images and Videos: Age Estimation and Drowsiness Detection.},
  author = {Salah Eddine Bekhouche},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:pyW8ca7W8N0C}
}

Abstract

Art style classification remains a formidable challenge in computational aesthetics due to the scarcity of expertly labeled datasets and the intricate, often nonlinear interplay of stylistic elements. While recent dual-teacher self-supervised frameworks reduce reliance on labeled data, their linear projection layers and localized focus struggle to model global compositional context and complex style-feature interactions. We enhance the dual-teacher knowledge distillation framework to address these limitations by replacing conventional MLP projection and prediction heads with Kolmogorov–Arnold Networks (KANs). Our approach retains complementary guidance from two teacher networks, one emphasizing localized texture and brushstroke patterns, the other capturing broader stylistic hierarchies while leveraging KANs’ spline-based activations to model nonlinear feature correlations with mathematical precision …

BibTeX

@misc{Sellam2025BeyondLinear,
  title  = {Beyond Linear Bottlenecks: Spline-Based Knowledge Distillation for Culturally Diverse Art Style Classification},
  author = {Abdellah Zakaria Sellam and Salah Eddine Bekhouche and Cosimo Distante and Abdelmalik Taleb-Ahmed},
  booktitle= {International Conference on Image Analysis and Processing},
  pages  = {401-412},
  publisher= {Springer Nature Switzerland},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:2P1L_qKh6hAC}
}

Abstract

Fine-grained object detection in challenging visual domains, such as vehicle damage assessment, presents a formidable challenge even for human experts to resolve reliably. While DiffusionDet has advanced the state-of-the-art through conditional denoising diffusion, its performance remains limited by local feature conditioning in context-dependent scenarios. We address this fundamental limitation by introducing Context-Aware Fusion (CAF), which leverages cross-attention mechanisms to integrate global scene context with local proposal features directly. The global context is generated using a separate dedicated encoder that captures comprehensive environmental information, enabling each object proposal to attend to scene-level understanding. Our framework significantly enhances the generative detection paradigm by enabling each object proposal to attend to comprehensive environmental information. Experimental results demonstrate an improvement over state-of-the-art models on the CarDD benchmark, establishing new performance benchmarks for context-aware object detection in fine-grained domains

BibTeX

@misc{Sellam2025CDiffDet,
  title  = {C-DiffDet+: Fusing Global Scene Context with Generative Denoising for High-Fidelity Car Damage Detection},
  author = {Abdellah Zakaria Sellam and Ilyes Benaissa and Salah Eddine Bekhouche and Abdenour Hadid and Vito Renó and Cosimo Distante},
  journal= {arXiv preprint arXiv:2509.00578},
  year   = {2025},
  eprint = {2509.00578},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:D03iK_w7-QYC}
}

Abstract

Fine-grained object detection in challenging visual domains, such as vehicle damage assessment, presents a formidable challenge even for human experts to resolve reliably. While DiffusionDet has advanced the state-of-the-art through conditional denoising diffusion, its performance remains limited by local feature conditioning in context-dependent scenarios. We address this fundamental limitation by introducing Context-Aware Fusion (CAF), which leverages cross-attention mechanisms to integrate global scene context with local proposal features directly. The global context is generated using a separate dedicated encoder that captures comprehensive environmental information, enabling each object proposal to attend to scene-level understanding. Our framework significantly enhances the generative detection paradigm by enabling each object proposal to attend to comprehensive environmental information. Experimental results demonstrate an improvement over state-of-the-art models on the CarDD benchmark, establishing new performance benchmarks for context-aware object detection in fine-grained domains

BibTeX

@misc{Sellam2025CDiffDet,
  title  = {C-DiffDet+: Fusing Global Scene Context with Generative Denoising for High-Fidelity Object Detection},
  author = {Abdellah Zakaria Sellam and Ilyes Benaissa and Salah Eddine Bekhouche and Abdenour Hadid and Vito Renó and Cosimo Distante},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:ZHo1McVdvXMC}
}

Abstract

Autonomous driving demands robust and real-time object detection to safely navigate in complex environments. While Convolutional neural network (CNN)-based detectors have been widely adopted, they face challenges such as limited receptive fields and inefficiencies in handling small or occluded objects. This paper presents a deformable Transformer based object detection framework designed to address these limitations. By leveraging deformable attention mechanisms, the model dynamically focuses on relevant spatial regions, significantly enhancing detection accuracy. Evaluated on the benchmark KITTI dataset, our proposed approach achieves an interesting mAP@50 of 96.6%, surpassing many state-of-the-art methods, at the cost of slower inference speed (7.0 FPS). The experimental results also demonstrate the framework’s superior precision and adaptability in autonomous driving scenarios. This …

BibTeX

@inproceedings{Kezzal2025DeformableTransforme,
  title  = {Deformable Transformer-Based Object Detection for Robust Perception in Autonomous Driving},
  author = {Chahira Kezzal and Selsabil Benderradji and Azeddine Benlamoudi and Salah Eddine Bekhouche and Abdel Taleb and Abdenour Hadid},
  booktitle= {2025 8th International Conference on Signal Processing and Information Security (ICSPIS)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:_xSYboBqXhAC}
}

Abstract

Accurate and real-time multi-object tracking (MOT) is essential for autonomous driving systems to ensure safe navigation and decision making in dynamic environments. This paper introduces a tracking-by-detection pipeline that integrates YOLOv11 a high speed, high-accuracy object detector with ByteTrack, a robust data association algorithm capable of lever-aging both high and low confidence detections. The proposed framework addresses key challenges in MOT such as frequent occlusions, fluctuating lighting, and dense traffic by combining efficient detection with motion-consistent identity tracking. Evaluated on the KITTI benchmark, our method demonstrates superior performance across multiple metrics, including HOTA, AssA, and MOTA, for both cars and pedestrians. Additionally, the system achieves an average runtime of 60.4 FPS, supporting its real-time applicability. The results confirm that the proposed …

BibTeX

@inproceedings{Benderradji2025EfficientReal,
  title  = {Efficient Real-Time Multi-Class Object Tracking with YOLO11 and ByteTrack in Real-World Driving Scenes},
  author = {Selsabil Benderradji and Chahira Kezzal and Azeddine Benlamoudi and Salah Eddine Bekhouche and Abdel Taleb and Abdenour Hadid},
  booktitle= {2025 8th International Conference on Signal Processing and Information Security (ICSPIS)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2025},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:bFI3QPDXJZMC}
}

Abstract

Pedestrian Attribute Recognition (PAR) involves predicting fine-grained attributes such as clothing color, gender, and accessories from pedestrian imagery, yet is hindered by severe class imbalance, intricate attribute co-dependencies, and domain shifts. We introduce VLM-PAR, a modular vision-language framework built on frozen SigLIP 2 multilingual encoders. By first aligning image and prompt embeddings via refining visual features through a compact cross-attention fusion, VLM-PAR achieves significant accuracy improvement on the highly imbalanced PA100K benchmark, setting a new state-of-the-art performance, while also delivering significant gains in mean accuracy across PETA and Market-1501 benchmarks. These results underscore the efficacy of integrating large-scale vision-language pretraining with targeted cross-modal refinement to overcome imbalance and generalization challenges in PAR.

BibTeX

@misc{Sellam2025VLMPAR,
  title  = {VLM-PAR: A Vision Language Model for Pedestrian Attribute Recognition},
  author = {Abdellah Zakaria Sellam and Salah Eddine Bekhouche and Fadi Dornaika and Cosimo Distante and Abdenour Hadid},
  journal= {arXiv preprint arXiv:2512.22217},
  year   = {2025},
  eprint = {2512.22217},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:EUQCXRtRnyEC}
}

Abstract

Over the last decade, the world has witnessed many breakthroughs in artificial intelligence, largely due to advances in deep learning technology. Notably, computer vision solutions have significantly contributed to these achievements. Human face analysis, a core area of computer vision, has gained considerable attention due to its wide applicability in fields such as law enforcement, social media, and marketing. However, existing methods for facial age estimation often struggle with accuracy due to limited feature extraction capabilities and inefficiencies in learning hierarchical representations. This paper introduces a novel framework to address these issues by proposing a Multi-Stage Deep Neural Network (MSDNN) architecture. The MSDNN architecture divides each CNN backbone into multiple stages, enabling more comprehensive feature extraction, thereby improving the accuracy of age predictions from facial images. Our framework demonstrates a significant performance improvement over traditional solutions, with its effectiveness validated through comparisons with the EfficientNet and MobileNetV3 architectures. The proposed MSDNN architecture achieves a notable decrease in Mean Absolute Error across three widely used public datasets (MORPH2, CACD, and AFAD) while maintaining a virtually identical parameter count compared to the initial backbone architectures. These results underscore the effectiveness and feasibility of our methodology in advancing the field of age estimation, showcasing it as a robust solution for enhancing the accuracy of age prediction algorithms.

BibTeX

@article{bekhouche2024facialage,
  title   = {Facial Age Estimation Using Multi-Stage Deep Neural Networks},
  author  = {Bekhouche, Salah Eddine and Benlamoudi, Azeddine and Dornaika, Fadi and Telli, Hichem and Bounab, Yazid},
  journal = {Electronics},
  volume  = {13},
  number  = {16},
  pages   = {3259},
  year    = {2024},
  doi     = {10.3390/electronics13163259}
}

Abstract

In recent years, natural language processing (NLP) has gained new territory beyond its traditional use in text mining applications. This paper shows the effectiveness of NLP techniques in assessing the apparent human personality from his/her video transcript, building a bridge between NLP and computer vision-based reasoning. In this paper, a new deep learning model using attention mechanism and bidirectional LSTM layers for estimating the Big-five personality traits is provided and then tested on ChatLearn video dataset. The robustness of the approach is then tested by generalizing the method to two other datasets (B5 corpus and Mypersonality datasets). Several empirical evaluations taking into account the various inputs of the data processing pipeline have been performed to yield optimal model parameters. The developed model is tested on the APA’2016 competition dataset from Chalearn V2 Challenge …

BibTeX

@article{Bounab2024Towardsjob,
  title  = {Towards job screening and personality traits estimation from video transcriptions},
  author = {Yazid Bounab and Mourad Oussalah and Nabil Arhab and Salah Bekhouche},
  journal= {Expert Systems with Applications},
  volume = {238},
  pages  = {122016},
  publisher= {Pergamon},
  year   = {2024},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:ns9cj8rnVeAC}
}

Abstract

The National Safety Council (NSC) has reported that each year, driver fatigue is responsible for 100,000 accidents, 71,000 injuries, and 1,550 fatalities, often manifesting as drowsiness. Most current systems fail to anticipate accidents beforehand, focusing primarily on external factors. This paper introduces a novel drowsiness detection framework aimed at reducing accidents caused by drivers dozing off behind the wheel and minimizing harm to individuals engaged in prolonged computer use. The proposed method leverages a Convolutional Neural Network (CNN) with a semi-supervised learning technique, setting it apart from existing approaches. The effectiveness of our approach is evaluated using the UTA-RLDD video dataset, and the results demonstrate that the proposed method outperforms state-of-the-art methods, achieving 99.98% accuracy.

BibTeX

@inproceedings{Rahmani2024AdvancingRoad,
  title  = {Advancing Road Safety: A CNN-Based Semi-Supervised Learning Approach for Drivers' Drowsiness Detection},
  author = {Chahrazad Rahmani and Azeddine Benlamoudi and Salah Eddine Bekhouche and Yazid Bounab and Djamel Samai and Fadi Dornaika},
  booktitle= {2024 1st International Conference on Electrical, Computer, Telecommunication and Energy Technologies (ECTE-Tech)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2024},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:J_g5lzvAfSwC}
}

BibTeX

@misc{NIU2024Devicesmethods,
  title  = {Devices, methods and computer programs for virtual eyeglasses try-on},
  author = {L NIU and MEN ZIGHEM and SE BEKHOUCHE},
  year   = {2024},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:pqnbT2bcN3wC}
}

Abstract

The objective of kinship verification is to assess whether two individuals are biologically related or not. Previous research has primarily focused on determining kinship from facial patterns and movements, voice, or human gait. In this paper, we explore for the first time in the literature the problem of kinship verification from text. This is a very timely topic given the emergence of generative artificial intelligence and large language models. Our main hypothesis is that two family members inherit and share subtitle textual features that can be seen in the way of writing. The subtle features can be related to genes, culture, experience etc. To address this problem, we propose a Siamese-based BERT transformer incorporating two novel modules, namely Attention and Fusion. The experiments show promising results on the role textual information in kinship verification.

BibTeX

@inproceedings{Bekhouche2024KinshipVerification,
  title  = {Kinship Verification from Text: Towards Discovering Subtitle Textual Features Shared by Family Members using Large Language Models},
  author = {Salah Eddine Bekhouche and Abdenour Hadid},
  booktitle= {2024 IEEE 6th International Conference on AI Circuits and Systems (AICAS)},
  pages  = {507-511},
  publisher= {IEEE},
  year   = {2024},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:RGFaLdJalmkC}
}

Research summary

This work presents a computer-vision framework for detecting driver drowsiness in video sequences through hybrid deep-feature selection. The method extracts representations from facial video data and selects complementary features to improve drowsiness recognition. The publisher does not expose the full abstract through its open metadata; follow the DOI for the authoritative abstract.

BibTeX

@article{bekhouche2022drowsiness,
  title   = {Driver Drowsiness Detection in Video Sequences Using Hybrid Selection of Deep Features},
  author  = {Bekhouche, Salah Eddine and Ruichek, Yassine and Dornaika, Fadi},
  journal = {Knowledge-Based Systems},
  volume  = {252},
  pages   = {109436},
  year    = {2022},
  doi     = {10.1016/j.knosys.2022.109436}
}

Abstract

Currently, face recognition technology is the most widely used method for verifying an individual’s identity. Nevertheless, it has increased in popularity, raising concerns about face presentation attacks, in which a photo or video of an authorized person’s face is used to obtain access to services. Based on a combination of background subtraction (BS) and convolutional neural network(s) (CNN), as well as an ensemble of classifiers, we propose an efficient and more robust face presentation attack detection algorithm. This algorithm includes a fully connected (FC) classifier with a majority vote (MV) algorithm, which uses different face presentation attack instruments (e.g., printed photo and replayed video). By including a majority vote to determine whether the input video is genuine or not, the proposed method significantly enhances the performance of the face anti-spoofing (FAS) system. For evaluation, we considered the MSU MFSD, REPLAY-ATTACK, and CASIA-FASD databases. The obtained results are very interesting and are much better than those obtained by state-of-the-art methods. For instance, on the REPLAY-ATTACK database, we were able to attain a half-total error rate (HTER) of 0.62% and an equal error rate (EER) of 0.58%. We attained an EER of 0% on both the CASIA-FASD and the MSU MFSD databases.

BibTeX

@article{Benlamoudi2022FacePresentation,
  title  = {Face Presentation Attack Detection Using Deep Background Subtraction},
  author = {Azeddine Benlamoudi and Salah Eddine Bekhouche and Maarouf Korichi and Khaled Bensid and Abdeldjalil Ouahabi and Abdenour Hadid and Abdelmalik Taleb-Ahmed},
  journal= {Sensors},
  volume = {22},
  number = {10},
  pages  = {3760},
  publisher= {MDPI},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:NMxIlDl6LWMC}
}

Abstract

Eye blink detection is a challenging problem that many researchers are working on because it has the potential to solve many facial analysis tasks, such as face anti-spoofing, driver drowsiness detection, and some health disorders. There have been few attempts to detect blinking in the wild scenario, while most of the work has been done under controlled conditions. Moreover, current learning approaches are designed to process sequences that contain only a single blink ignoring the case of the presence of multiple eye blinks. In this work, we propose a fast framework for eye blink detection and eye blink verification that can effectively extract multiple blinks from image sequences considering several challenges such as lighting changes, variety of poses, and change in appearance. The proposed framework employs fast landmarks detector to extract multiple facial key points including the ones that identify the eye …

BibTeX

@article{Bekhouche2022SpatiotemporalCNN,
  title  = {Spatiotemporal CNN with Pyramid Bottleneck Blocks: Application to eye blinking detection},
  author = {SE Bekhouche and I Kajo and Y Ruichek and F Dornaika},
  journal= {Neural Networks},
  volume = {152},
  pages  = {150-159},
  publisher= {Pergamon},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:blknAaTinKkC}
}

Abstract

In this paper, we propose a framework for improving Background Subtraction techniques. This framework is based on two types of data, RGB and Depth. Our study stands for obtaining preliminary results of the background segmentation using Depth and RGB channels independently, then using an algorithm to fuse them to create the final results. The experiments on the SBM-RGBD dataset using four methods: ViBe, LOBSTER, SuBSENSE, and PAWCS proved that the proposed framework achieves an impressive performance compared to the original RGB-based techniques from the state-of-the-art.

BibTeX

@inproceedings{Houhou2022ImprovingViBe,
  title  = {Improving ViBe-based Background Subtraction Techniques Using RGBD Information},
  author = {Ihssane Houhou and Athmane Zitouni and Yassine Ruichek and Salah Eddine Bekhouche and Abdelmalik Taleb-Ahmed},
  booktitle= {2022 7th International Conference on Image and Signal Processing and their Applications (ISPA)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:YFjsv_pBGBYC}
}

Abstract

This paper proposes a novel deep learning model called deep multi-scale network (DMSN) for background subtraction. This convolutional neural network is built to use RGB color channels and Depth maps as inputs with which it can fuse semantic and spatial information. In comparison with previous deep learning background subtraction techniques that lack information due to its use of only RGB channels, our RGBD version is able to overcome most of the drawbacks, especially in some particular kinds of challenges. Further, this paper introduces a new protocol for the SBM-RGBD dataset, concerning scene-independent evaluation, dedicated to Deep Learning methods to set up a competitive platform that includes more challenging situations. The proposed method proved its efficiency in solving the background subtraction in complex situations at different levels. The experimental results verify that the proposed …

BibTeX

@article{Houhou2022RGBDdeep,
  title  = {RGBD deep multi-scale network for background subtraction},
  author = {Ihssane Houhou and Athmane Zitouni and Yassine Ruichek and Salah Eddine Bekhouche and Mohamed Kas and Abdelmalik Taleb-Ahmed},
  journal= {International Journal of Multimedia Information Retrieval},
  volume = {11},
  number = {3},
  pages  = {395-407},
  publisher= {Springer London},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:hMod-77fHWUC}
}

Abstract

Metric learning has attracted wide attention in face and kinship verification and a number of such algorithms have been presented over the past few years, this system has a number of applications such as organizing collections of images and recognizing resemblances among humans and finding of missing children. In this work , we propose a novel approach based on the Weber Local Descriptor (WLD) with color spaces, and the Multi-Lavel (ML) representation, Moreover, the use of Rank features (TTest) to reduce the number of features and the support vector machine (SVM) for the kinship classification. Our approach consists of six stages which are : (1) Face preprocessing (2) applied the color spaces (3) features extraction using WLD , with face representation , (4) pair features representation, (5) features selection and (6) classification using SVM. The proposed approach istested and analyzed on five publicly …

BibTeX

@inproceedings{Bouakal2022KinshipVerification,
  title  = {Kinship Verification System based on the Color Spaces Analysis},
  author = {Nour El Houda Bouakal and Messaoud Hettiri and Abdelhakim Chergui and Abdelkrim Ouafi and Azeddine Benlamoudi and Salah Eddine Bekhouche},
  booktitle= {2022 7th International Conference on Image and Signal Processing and their Applications (ISPA)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:BqipwSGYUEgC}
}

Abstract

Introduction The 3D integrated circuit technology, which smooths out the massive increase in transistors on a chip by stacking numerous silicon layers vertically, is quickly becoming a revolutionary technology. Thermal issues are more relevant for 3D Network-on-Chip (NoC) systems than their 2D counterparts. Methods This paper presents a novel Vertically-Partially-Connected 3D-Network on-chip architecture that reduces the total length of interconnects and reduces the number of 3D routers. We also present an efficient XYZ routing technique for thermal management. The proposed algorithm distributes traffic based on the number of layers and congestion to achieve chip heat balancing, avoid high peak temperatures, improve average packet latency, and extend chip service life. Results Simulation results showed that the routing technique reduces the peak temperature of the chip by an average of 17 °C …

BibTeX

@article{Benhaoues2022EffectiveRouting,
  title  = {Effective Routing Algorithm for Thermal Management in Vertically-Partially-Connected 3D-network on Chip},
  author = {Atef Benhaoues and Abdelhalim Rabehi and Salah E Bekhouche and El-Bay Bourennane},
  journal= {Recent Advances in Electrical \& Electronic Engineering (Formerly Recent Patents on Electrical \& Electronic Engineering)},
  volume = {15},
  number = {8},
  pages  = {664-678},
  publisher= {Bentham Science Publishers},
  year   = {2022},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:lSLTfruPkqcC}
}

Abstract

Recently, automatic personality analysis is becoming an interesting topic for computer vision. Many attempts have been proposed to solve this problem using time-based sequence information. In this paper, we present a new framework for estimating the Big-Five personality traits and job candidate screening variable from video sequences. The framework consists of two parts: (1) the use of Pyramid Multi-level (PML) to extract raw facial textures at different scales and levels; (2) the extension of the Covariance Descriptor (COV) to fuse different local texture features of the face image such as Local Binary Patterns (LBP), Local Directional Pattern (LDP), Binarized Statistical Image Features (BSIF), and Local Phase Quantization (LPQ). Therefore, the COV descriptor uses the textures of PML face parts to generate rich low-level face features that are encoded using concatenation of all PML blocks in a feature vector. Finally, the entire video sequence is represented by aggregating these frame vectors and extracting the most relevant features. The exploratory results on the ChaLearn LAP APA2016 dataset compare well with state-of-the-art methods including deep learning-based methods.

BibTeX

@article{Telli2021ANovel,
  title  = {A Novel Multi-Level Pyramid Co-Variance Operators for Estimation of Personality Traits and Job Screening Scores},
  author = {Hichem Telli and Salim Sbaa and Salah Eddine Bekhouche and Fadi Dornaika and Abdelmalik Taleb-Ahmed and Miguel Bordallo López},
  journal= {Traitement du Signal},
  volume = {38},
  number = {3},
  pages  = {539-546},
  year   = {2021},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:maZDTaKrznsC}
}

Abstract

In recent times, human facial age estimation topic attracted a lot of attention due to its ability to improve biometrics systems. Recently, several applications that exploit demographic attributes have emerged. These applications include: access control, re-identification in surveillance videos, integrity of face images in social media, intelligent advertising, human–computer interaction, and law enforcement. In this chapter, we present a novel approach for human facial age estimation in facial images. The proposed approach consists of the following three main stages: (1) face preprocessing; (2) feature extraction (two different kinds of features are studied: handcrafted and deep features); (3) feeding the obtained features to a linear regressor. Also, we investigate the strength and weakness of handcrafted and deep features for facial age estimation. Experiments are conducted on three public databases (FG-NET, PAL and …

BibTeX

@misc{Bekhouche2021HumanFacial,
  title  = {Human Facial Age Estimation: Handcrafted Features Versus Deep Features},
  author = {Salah Eddine Bekhouche and Fadi Dornaika and Abdelkrim Ouafi and Abdelmalik Taleb-Ahmed},
  booktitle= {Emerging Technologies in Biomedical Engineering and Sustainable TeleMedicine},
  pages  = {31-37},
  publisher= {Springer, Cham},
  year   = {2021},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:M3NEmzRMIkIC}
}

Abstract

Recent works have shown that deep Convolutional Neural Networks (CNNs) can be very effective for image-based age estimation. However, the proposed approaches significantly vary, and there are still some open problems. Almost all deep regression networks for age estimation have exploited the Mean Square Error loss only. These deep networks have not considered the influence of aberrant and outlier observations on the final model. In this letter, we introduce the use of robust loss functions in order to learn deep regression networks for age estimation. More precisely, we explore the use of two robust regression functions: (i) the L1 norm error, and (ii) the adaptive loss function that retains the advantages of the L1 and L2 norms. Experimental results obtained on four public databases demonstrate that learning a deep CNN with robust losses can improve age estimation.

BibTeX

@article{dornaika2020robust,
  title   = {Robust Regression with Deep CNNs for Facial Age Estimation: An Empirical Study},
  author  = {Dornaika, Fadi and Bekhouche, Salah Eddine and Arganda-Carreras, Ignacio},
  journal = {Expert Systems with Applications},
  volume  = {141},
  pages   = {112942},
  year    = {2020},
  doi     = {10.1016/j.eswa.2019.112942}
}

Abstract

The use of facial images in the kinship verification is a challenging research problem in soft biometrics and computer vision. In our work, we present a kinship verification system that starts with pair of facial images of the child and parent, then as a final result is determine whether two persons have a kin relation or not. our approach contains five steps as follows: (i) the face preprocessing step to get aligned and cropped facial images of the pair (ii), extracting deep features based on the deep learning model called Visual Geometry Group (VGG) Face, (iii) applying our proposed pair feature representation function alongside with a features normalization, (iv) the use of Fisher Score (FS) to select the best discriminative features, (v) decide whether there is a kinship or not based on the Support Vector Machine (SVM) classifier. We conducted several experiments to demonstrate the effectiveness of our approach that we tested on five benchmark databases (Cornell KinFace, UB KinFace, Familly101, KinFace W-I, and KinFace W-II). Our results indicate that our system is robust compared to other existing approaches.

BibTeX

@article{Chergui2020Kinshipverification,
  title  = {Kinship verification through facial images using cnn-based features},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sébastien Mavromatis and Salah Eddine Bekhouche and Mohamed Lashab and Jean Sequeira},
  journal= {Traitement du Signal},
  volume = {37},
  number = {1},
  pages  = {1-8},
  year   = {2020},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:e5wmG9Sq2KIC}
}

Abstract

In recent times, the topic of human facial age estimation attracted much attention. This is due to its ability to improve biometrics systems. Recently, several applications that are based on the demographic attributes estimation have been developed. These include law enforcement, re-identification in videos, planed marketing, intelligent advertising, social media, and human-computer interaction. The main contributions of the paper are as follows. Firstly, it extends some handcrafted models that are based on the Pyramid Multi Level (PML) face representation. Secondly, it evaluates the performance of two different kinds of features that are handcrafted and deep features. It compares handcrafted and deep features in terms of accuracy and computational complexity. The proposed scheme of study includes the following three main steps: 1) face preprocessing; 2) feature extraction (two different kinds of features are studied …

BibTeX

@article{Bekhouche2020Acomparative,
  title  = {A comparative study of human facial age estimation: handcrafted features vs. deep features},
  author = {SE Bekhouche and F Dornaika and A Benlamoudi and A Ouafi and A Taleb-Ahmed},
  journal= {Multimedia Tools and Applications},
  volume = {79},
  number = {35},
  pages  = {26605-26622},
  publisher= {Springer US},
  year   = {2020},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:r0BpntZqJG4C}
}

Abstract

Kinship verification aims to determine whether two persons are kin related or not. This is an emerging topic in computer vision due to its practical potential applications such as family album management. Most of previous works are based on checking kinship from face patterns and more recently from voices. We provide in this paper the first investigation in the literature on kinship verification from gait. The main purpose is to study whether family members do share some gait patterns. As this is a new topic, we started by collecting a new dataset for kinship verification from human gait containing several pairs of video sequences of celebrities and their relatives. The database will be released to the research community for research purposes. Along with the database, we provide results using baseline methods using silhouette and video based analysis. Moreover, we also propose a two-stream 3DCNN to tackle the …

BibTeX

@inproceedings{Bekhouche2020KinshipVerification,
  title  = {Kinship Verification From Gait?},
  author = {Salah Eddine Bekhouche and Abdelhakim Chergui and Abdenour Hadid and Yassine Ruichek},
  booktitle= {2020 IEEE International Conference on Image Processing (ICIP)},
  pages  = {2950-2954},
  publisher= {IEEE},
  year   = {2020},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:iH-uZ7U-co4C}
}

Abstract

We address kinship verification, which is a challenging problem in computer vision and pattern discovery. It has several applications, such as organizing photoalbums, recognizing resemblances among humans, and finding missing children. We present a system for facial kinship verification based on several kinds of texture descriptors (local binary patterns, local ternary patterns, local directional patterns, local phase quantization, and binarized statistical image features) with pyramid multilevel (PML) face representation for feature extraction along with our proposed paired feature representation and our proposed robust feature selection to reduce the number of features. The proposed approach consists of the following three main stages: (1) face preprocessing, (2) feature extraction and selection, and (3) kinship verification. Extensive experiments are conducted on five publicly available databases (Cornell, UB …

BibTeX

@article{Chergui2020Kinshipverification,
  title  = {Kinship verification through facial images using multiscale and multilevel handcrafted features},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sebastien Mavromatis and Salah Eddine Bekhouche and Jean Sequeira and Fadi Dornaika},
  journal= {Journal of Electronic Imaging},
  volume = {29},
  number = {2},
  pages  = {023017-023017},
  publisher= {Society of Photo-Optical Instrumentation Engineers},
  year   = {2020},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:j3f4tGmQtD8C}
}

Abstract

The kinship verification through facial images is an active research topic due to its potential applications. In this paper, we propose an approach which takes two images as an input then give kinship result (kinship / No-kinship) as an output. The approach contains five steps which are: (1) face preprocessing, (2) deep features extraction, (3) pair features representation and normalization, (4) features selection, (5) kinship verification. Experiments are conducted on five public databases (Cornell KinFace, UB Kin database, Familly, KinFace-I, and KinFace-II). The experimental results shows that our approach is comparable with existed approaches.

BibTeX

@inproceedings{Chergui2019DeepFeatures,
  title  = {Deep Features for Kinship Verification from Facial Images},
  author = {Abdelhakim Chergui and Salim Ouchtati and Jean Sequeira and Salah Eddine Bekhouche and Fares Bougourzi and Hichem Telli},
  booktitle= {2019 International Conference on Advanced Systems and Emergent Technologies (IC_ASET)},
  pages  = {64-67},
  publisher= {IEEE},
  year   = {2019},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:-f6ydRqryjwC}
}

Abstract

The kinship verification through facial images is ana ctive research topic due to its potential applications. In this paper, we propose an approach which takes two images as input then give kinship result (kinship / No-kinship) as an output. our approach based on the deep learning model (ResNet) for the feature extraction step, alongside with our proposed pair feature representation function and RankFeatures (Ttest) for feature selection to reduce the number of features finally we use the SVM classifier for the decision of kinship verification. The approach contains three steps which are: (1) face preprocessing, (2) deep features extraction and pair features representation (3) Classification. Experiments are conducted on five public databases. The experimental results show that our approach is comparable with existed approaches.

BibTeX

@inproceedings{Chergui2019InvestigatingDeep,
  title  = {Investigating Deep CNNs Models Applied in Kinship Verification through Facial Images},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sebastien Mavromatis and Salah Eddine Bekhouche and Jean Sequeira},
  booktitle= {2019 5th International Conference on Frontiers of Signal Processing (ICFSP)},
  pages  = {82-87},
  publisher= {IEEE},
  year   = {2019},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:mB3voiENLucC}
}

Abstract

Kinship verification is a challenging problem that recently attracted much interest in computer vision, this system has a number of applications such as organizing large collections of images and recognizing resemblances among humans and search for lost people. In this work, we propose a new method based on different descriptors mixed such as (LBP, LPQ, BSIF), and the Multi-Block (MB) representation. and we investigate the effect of different features representation for kinship verification, Moreover, the use of TTest to reduce the number of features and the support vector machine (SVM) for the kinship classification. Our approach consists of five stages : (1) features extraction , (2) face representation (3) features representation, (4) features selection and (5) classification. Our approach is tested on five datasets (Cornell, UB Kin Face, Familly 101, KinFac W-I and W-II). Our results are good comparable with other …

BibTeX

@inproceedings{Chergui2019KinshipVerification,
  title  = {Kinship Verification using Mixed Descriptors and Multi Block Face Representation},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sebastien Mavromatis and Salah Eddine Bekhouche and Jean Sequeira and Houssem Zerrari},
  booktitle= {2019 International Conference on Networking and Advanced Systems (ICNAS)},
  pages  = {1-6},
  publisher= {IEEE},
  year   = {2019},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:IWHjjKOFINEC}
}

Abstract

This paper deals with a two-stage method for the sizing of two standalone solar lighting systems. Where, each of these systems uses a solar panel, a battery backup, a DC lamp and a charge controller. These remotely controlled systems are intended to be used in public places and streets of Biskra and Skikda for replacing the traditional grid-powered lighting systems. The first stage is a pre-sizing step in which the consumed daily energy, the system autonomy, the peak sun hours and the system losses are considered for the sizing of the battery and the photovoltaic generator. In the second stage, the PVSYST software is used to evaluate and optimize the size of the PV generator by simulating the systems over a whole year.The obtained results show the importance of the tilt angle choice in the production production and the efficiency of the solar panel. Indeed, as the value of this angle is approximately equal to the …

BibTeX

@inproceedings{Zerari2019Atwo,
  title  = {A two-stage sizing method of standalone solar lighting systems},
  author = {Houssam Zerari and Lotfi Messikh and Abdelhakim Dendouga and Abdellah Kouzou and Salah Eddine Bekhouche and Salim Ouchtati and Abdelhakim Chergui},
  booktitle= {2019 4th International Conference on Power Electronics and their Applications (ICPEA)},
  pages  = {1-5},
  publisher= {IEEE},
  year   = {2019},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:hFOr9nPyWt4C}
}

BibTeX

@article{Chergui2019Fusinghandcrafted,
  title  = {Fusing handcrafted features for facial kinship verification systems},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sébastien Mavromatis and Salah Eddine Bekhouche and Jean Sequeira and Houssem Zerrari},
  journal= {International Journal of Strategic Information Technology and Applications (IJSITA)},
  year   = {2019},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:vV6vV6tmYwMC}
}

Abstract

The automatic verification of kinship is a challenging problem that recently attracted much interest in computer vision, the kinship verification has become an active research field due to its potential applications such as organizing photo albums and images annotation, recognizing resemblances among humans and finding of missing children. In this paper, we propose an approach which takes two images as an input then give kinship result (kinship/non-kinship) as an output. This approach based on the Local Phase Quantization (LPQ) and Local directional pattern (LDP) features descriptors and the ML (Multi-Level) representation for the kinship verification from facial images, this work consists six stages which are:(i) face preprocessing,(ii) features extraction,(iii) face representation (iv) pair features representation and normalization,(v) features selection and (vi) kinship verification. Experiments are conducted on four public databases (Cornell KinFace, UB Kin database, KinFace-I, and KinFace-II). The obtained results are good compared with state-of-the-art approaches.

BibTeX

@misc{Chergui2018LPQand,
  title  = {LPQ and LDP Descriptors with ML Representation For Kinship verification},
  author = {Abdelhakim Chergui and Salim Ouchtati and Hichem Telli and Fares Bougourzi and Salah Eddine Bekhouche},
  year   = {2018},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:7PzlFSSx8tAC}
}

Abstract

In recent years, the kinship verification has got more intention due to its potential applications like images annotation, organizing the photo albums and recognizing similarities among humans. In this paper, we propose a new approach based on the Local Binary Pattern (LBP) and Binarized Statistical Image features (BSIF) descriptors and the PML (Pyramid Multi-Level) representation for the kinship verification from the static images. In particular, we investigate the best parameters of each method (LBP, BSIF and PML) for the kinship verification, then conclude the suitable combination. The approach consists three main stages which are: (1) Face preprocessing,(2) features extraction, and (3) classification and decision (kin or non kin). The proposed approach is tested and analyzed on four public databases (Cornell KinFace, UB Kin database, KinFace-I, and KinFace-II). The obtained results of our approach have …

BibTeX

@inproceedings{Chergui2018Kinshipverification,
  title  = {Kinship verification using BSIF and LBP},
  author = {Abdelhakim Chergui and Salim Ouchtati and Jean Sequeira and Salah Eddine Bekhouche and Fares Bougourzi},
  booktitle= {2018 International Conference on Signal, Image, Vision and their Applications (SIVA)},
  pages  = {1-5},
  publisher= {IEEE},
  year   = {2018},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:L8Ckcad2t8MC}
}

BibTeX

@article{Chergui2018Discriminantanalysis,
  title  = {Discriminant analysis for facial verification using color images},
  author = {A Chergui and S Ouchtati and J Sequeira and SE Bekhouche and F Bougourzi},
  journal= {Proc. First Int. Conf. Electr. Eng},
  year   = {2018},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:RHpTSmoSYBkC}
}

BibTeX

@article{Chergui2018Robustkinship,
  title  = {Robust kinship verification using local descriptors},
  author = {A Chergui and S Ouchtati and J Sequeira and SE Bekhouche and Hichem Telli},
  journal= {Proc. Third Int. Conf. Adv. Technol. and Electr. Eng},
  year   = {2018},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:4JMBOYKVnBMC}
}

Abstract

Tiivistelmä Abstract In recent years, software-based face presentation attack detection (PAD) methods have seen a great progress. However, most existing schemes are not able to generalize well in more realistic conditions. The objective of this competition is to evaluate and compare the generalization performances of mobile face PAD techniques under some real-world variations, including unseen input sensors, presentation attack instruments (PAI) and illumination conditions, on a larger scale OULU-NPU dataset using its standard evaluation protocols and metrics. Thirteen teams from academic and industrial institutions across the world participated in this competition. This time typical liveness detection based on physiological signs of life was totally discarded. Instead, every submitted system relies practically on some sort of feature representation extracted from the face and/or background regions using hand-crafted, learned or hybrid descriptors. Interesting results and findings are presented and discussed in this paper.

BibTeX

@inproceedings{Boulkenafet2017Acompetition,
  title  = {A competition on generalized software-based face presentation attack detection in mobile scenarios},
  author = {Z Boulkenafet and J Komulainen and Z Akhtar and A Benlamoudi and D Samai and SE Bekhouche and A Ouafi and F Dornaika and A Taleb-Ahmed and L Qin and F Peng and LB Zhang and M Long and S Bhilare and V Kanhangad and A Costa-Pazo and E Vazquez-Fernandez and D Pérez-Cabo and JJ Moreira-Pésrez and D González-Jiménez and A Mohammadi and S Bhattacharjee and S Marcel and S Volkova and Y Tang and N Abe and L Li and X Feng and Z Xia and X Jiang and S Liu and R Shao and PC Yuen and W Almeida and F Andaló and R Padilha and G Bertocco and W Dias and J Wainer and R Torres and A Rocha and MA Angeloni and G Folego and A Godoy and A Hadid},
  booktitle= {Proc. Int. Joint Conf. Biometrics},
  pages  = {688-696},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:QIV2ME_5wuYC}
}

Abstract

In this paper, we propose a novel approach for estimating the Big Five personality traits and the job candidate screening attribute through facial videos. At running time, the proposed system feeds the Pyramid Multi-Level (PML) texture features extracted from the whole video sequence to 5 Support Vector Regressors in order to estimate the personality traits. These estimated five scores are then considered as new input features to the interview score regressor. The latter is given by a Gaussian Process Regression (GPR). The experimental results on ChaLearn LAP APA2016 dataset achieve good performance. Furthermore, they demonstrate that the computational cost of both the training and the testing of the proposed framework are very competitive in terms of accuracy and computational cost.

BibTeX

@inproceedings{Bekhouche2017Personalitytraits,
  title  = {Personality traits and job candidate screening via analyzing facial videos},
  author = {Salah Eddine Bekhouche and Fadi Dornaika and Abdelkrim Ouafi and Abdelmalik Taleb-Ahmed},
  booktitle= {Proceedings of the IEEE conference on computer vision and pattern recognition workshops},
  pages  = {10-13},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:Wp0gIr-vW9MC}
}

Abstract

We present a novel learning system for human demographic estimation in which the ethnicity, gender and age attributes are estimated from facial images. The proposed approach consists of the following three main stages: 1) face alignment and preprocessing; 2) constructing a Pyramid Multi-Level face representation from which the local features are extracted from the blocks of the whole pyramid; 3) feeding the obtained features to an hierarchical estimator having three layers. Due to the fact that ethnicity is by far the easiest attribute to estimate, the adopted hierarchy is as follows. The first layer predicts ethnicity of the input face. Based on that prediction, the second layer estimates the gender using the corresponding gender classifier. Based on the predicted ethnicity and gender, the age is finally estimated using the corresponding regressor. Experiments are conducted on five public databases (MORPH II, PAL, IoG …

BibTeX

@article{Bekhouche2017Pyramidmulti,
  title  = {Pyramid multi-level features for facial demographic estimation},
  author = {Salah Eddine Bekhouche and Abdelkrim Ouafi and Fadi Dornaika and Abdelmalik Taleb-Ahmed and Abdenour Hadid},
  journal= {Expert Systems with Applications},
  volume = {80},
  pages  = {297-310},
  publisher= {Pergamon},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:4TOpqqG69KYC}
}

Abstract

Soft biometrics topic attracted a lot of attention recently due to its ability to improve biometrics systems. It has a lot of traits which can be used in biometrics. Some of these traits is most popular among the other traits. These traits are called demographic traits (ie. age, gender, and ethnicity). It belongs to facial soft biometrics traits. Recently, several applications that exploit demographic attributes have emerged. These applications include: access control, reidentification in surveillance videos, integrity of face images in social media, intelligent advertising, human-computer interaction, and law enforcement. In this dissertation, facial demographic estimation through facial images is studied. Starting with the existing techniques like Deep Learning-based approaches, Image-Based approaches, and Anthropometrics-based approaches. Also, the databases used for age estimation, gender classification or ethnicity classification are exploited. Moreover, the different evaluation terms are mentioned. Ending with the proposed approach and the results on different databases. The proposed approach consists of the following three main stages: 1) face alignment and preprocessing; 2) feature extraction and selection; 3) demographic estimation. The purpose of face alignment is to localize faces in images, rectify the 2D or 3D pose of each face and crop the region of interest. This preprocessing stage is important since the subsequent stages depend on it and since it can affect the final performance of the system. The processing stage can be challenging since it should overcome many variations that may appear in the face image. Feature extraction and selection …

BibTeX

@misc{BEKHOUCHE2017FacialSoft,
  title  = {Facial Soft Biometrics: Extracting demographic traits},
  author = {Salah Eddine BEKHOUCHE},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:mVmsd5A6BfQC}
}

Abstract

The aim of this work is to investigate global and local image descriptors impact on facial gender classification by carrying out an independent comparative study among several texture descriptors algorithms. In this paper, we consider three global descriptors namely, Gray-Level Co-Occurrence Matrix (GLCM), Gabor Wavelet Transform (GWT) and Autocorrelation Function (ACF). On the other hand, we consider four local image descriptors called, Local Binary Patterns (LBP), Local Directional Pattern (LDP), Local Phase Quantization (LPQ) and Binarized Statistical Image Features (BSIF). The experimental comparison proofs that the local image descriptors are more efficient than the global ones in facial gender classification. All the experiments conducted on the Image of Groups (IoG) database.

BibTeX

@misc{Bougourzi2017Acomparative,
  title  = {A comparative study on textures descriptors in facial gender classification},
  author = {F Bougourzi and SE Bekhouche and ME Zighem and A Benlamoudi and A Ouafi and A Taleb-Ahmed},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:M3ejUd6NZC8C}
}

Abstract

The Face recognition applications are the used way of authentication identity verification of mobile payment. This popularity of face recognition is easy to raise concerns about face spoof attacks; use photo or video of an authorized person’s face to access to facilities or services. We propose an efficient and more robust algorithm for face spoof detection based-on combination between MLLBP and MLBSIF. An ensemble classifier, consisting of Lib-SVM classifiers using different face spoof attacks (eg, printed photo and replayed video) to trained our model which is used to distinguish between genuine and spoof faces. We tested our approach on CASIA FASD database. Our proposed approach conduct a good result compared with the state of art.

BibTeX

@misc{Benlamoudi2017FaceAnti,
  title  = {Face Anti-Spoofing combining MLLBP and MLBSIF},
  author = {A Benlamoudi and ME Zighem and F Bougourzi},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:4DMP91E08xMC}
}

Abstract

Automatic age estimation via faces is one of the dominant topics nowadays. In this paper, we proposed a novel approach for age estimation from RGB color face image. Despite the large number of methods for age estimation, their performance remains unsatisfactory for RGB images. The idea is to extract features from three color band instead of extract features from gray-scale. The experimental results show the superiority of our approach compared to that of the state-of-the-art methods that use the PAL dataset.

BibTeX

@misc{Zighem2017AgeEstimation,
  title  = {Age Estimation Based on Color Facial Texture},
  author = {ME Zighem and A Ouafi and SE Bekhouche and A Benlamoudi and A Taleb-Ahmed},
  year   = {2017},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:aqlVkmm33-oC}
}

Abstract

Human face aging is irreversible process causing changes in human face characteristics such us hair whitening, muscles drop and wrinkles. Due to the importance of human face aging in biometrics systems, age estimation became an attractive area for researchers. This paper presents a novel method to estimate the age from face images, using binarized statistical image features (BSIF) and local binary patterns (LBP)histograms as features performed by support vector regression (SVR) and kernel ridge regression (KRR). We applied our method on FG-NET and PAL datasets. Our proposed method has shown superiority to that of the state-of-the-art methods when using the whole PAL database.

BibTeX

@misc{Bekhouche2016Facialage,
  title  = {Facial age estimation using BSIF and LBP},
  author = {Salah Eddine Bekhouche and Abdelkrim Ouafi and Abdelmalik Taleb-Ahmed and Abdenour Hadid and Azeddine Benlamoudi},
  journal= {arXiv preprint arXiv:1601.01876},
  year   = {2016},
  eprint = {1601.01876},
  archivePrefix = {arXiv},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:3fE2CSJIrl8C}
}

Abstract

Facial demographic classification is an attractive topic in computer vision. Attributes such as age and gender can be used in many real life application such as face recognition and internet safety for minors. In this paper, we present a novel approach for age estimation and gender classification under uncontrolled conditions following the standard protocols for fair comparaison. Our proposed approach is based on Multi Level Local Phase Quantization (ML-LPQ) features which are extracted from normalized face images. Two different Support Vector Machines (SVM) models are used to predict the age group and the gender of a person. The experimental results on the benchmark Image of Groups dataset showed the superiority of our approach compared to the state-of-the-art.

BibTeX

@inproceedings{Bekhouche2015Facialage,
  title  = {Facial age estimation and gender classification using multi level local phase quantization},
  author = {Salah Eddine Bekhouche and Abdelkrim Ouafi and Azeddine Benlamoudi and Abdelmalik Taleb-Ahmed and Abdenour Hadid},
  booktitle= {2015 3rd International Conference on Control, Engineering \& Information Technology (CEIT)},
  pages  = {1-4},
  publisher= {IEEE},
  year   = {2015},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:kNdYIx-mwKoC}
}

Abstract

Todays biometric systems are vulnerable to spoof attacks made by non-real faces. The problem is when a person shows in front of camera a print photo or a picture from cell phone. We study in this paper an anti-spoofing solution for distinguishing between 'live' and 'fake' faces. In our approach we used overlapping block LBP operator to extract features in each region of the image. To reduce the features we used Fisher-Score. Finally, we used a nonlinear Support Vector Machine (SVM) classifier with kernel function for determining whether the input image corresponds to a live face or not. Our experimental analysis on a publicly available NUAA and CASIA face anti-spoofing databases following the standard protocols showed good results.

BibTeX

@inproceedings{Benlamoudi2015Facespoofing,
  title  = {Face spoofing detection using local binary patterns and Fisher score},
  author = {Azeddine Benlamoudi and Djamel Samai and Abdelkrim Ouafi and Salah Eddine Bekhouche and Abdelmalik Taleb-Ahmed and Abdenour Hadid},
  booktitle= {2015 3rd International Conference on Control, Engineering \& Information Technology (CEIT)},
  pages  = {1-5},
  publisher= {IEEE},
  year   = {2015},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:ULOm3_A8WrAC}
}

Abstract

Biometric technologies are becoming the foundation of an extensive array of highly secure identification and verification solution. Unfortunately, biometric systems are vulnerable to attacks made by persons showings photo, video or mask to spoof the real identity. In this paper we study a solution for those problems. We try to make solution to face spoofing for distinguishing between real face and fake one. Our approach called Multi-Level Local Phase Quantization (ML-LPQ) is focused in Local Phase Quantization (LPQ) descriptor for extracting features on face region of interest. In our approach, we use three levels for the LPQ descriptor to extract features and LibSVM for classification. Our experimental analysis on a publicly available CASIA face anti-spoofing database give us good result compared to other approaches using the same protocol.

BibTeX

@article{Benlamoudi2015Facespoofing,
  title  = {Face spoofing detection using multi-level local phase quantization (ML-LPQ)},
  author = {A Benlamoudi and D Samai and A Ouafi and SE Bekhouche and A Taleb-Ahmed and A Hadid},
  journal= {Proc. of the First Int. Conf. on Automatic Control, Telecommunication and signals ICATS15},
  year   = {2015},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:YOwf2qJgpHMC}
}

Abstract

Automatic age estimation and gender classification through facial images are attractive topics in computer vision. They can be used in many real-life applications such as face recognition and internet safety for minors. In this paper, we present a novel approach for age estimation and gender classification under uncontrolled conditions following the standard protocols for fair comparaison. Our proposed approach is based on Multi Level Local Binary Pattern (ML-LBP) features which are extracted from normalized face images. Two different Support Vector Machines (SVM) models are used to predict the age group and the gender of a person. The experimental results on benchmark Image of Groups dataset showed the superiority of our approach compared to that of the state-ofthe- art methods.

BibTeX

@misc{Bekhouche2015Automaticage,
  title  = {Automatic age estimation and gender classification in the wild},
  author = {SE Bekhouche and A Ouafi and A Benlamoudi and A Taleb-Ahmed and A Hadid},
  year   = {2015},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:Zph67rFs4hoC}
}

Abstract

Besides the recognition task, todays biometric systems need to cope with additional problem: spoofing attacks, like presenting a photo of a person (client) to camera. We study in this paper an anti-spoofing solution for distinguishing between’live’and’fake’faces. In our approach we focused in face detection using Viola-Jones algorithm and Active Shape Models with Stasm for locating landmarks. Then, we apply Local Binary Patterns (LBP) operator to extract the features in each region of the image. Finally, we use a nonlinear Support Vector Machine (SVM) classifier with kernel function for determining whether the input image corresponds to a live face or not. Our experimental analysis on a publicly available database NUAA, showed excellent results compared to existing methods.

BibTeX

@article{Benlamoudi2015Facespoofing,
  title  = {Face spoofing detection from single images using active shape models with stasm and lbp},
  author = {Azeddine Benlamoudi and Djamel Samai and Abdelkrim Ouafi and Abdelmalik Taleb-Ahmed and Salah Eddine Bekhouche and Abdenour Hadid},
  journal= {Proceeding of the Troisime Conference Internationale Sur La Vision Artificielle CVA 2015},
  year   = {2015},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:KlAtU1dfN6UC}
}

BibTeX

@misc{Bekhouche2013Ralisation,
  title  = {Réalisation d’un oscilloscope numérique portatif à affichage sur écran PC via le port USB},
  author = {Salah Eddine Bekhouche},
  year   = {2013},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:ZeXyd9-uunAC}
}

Abstract

The use of facial images in the kinship verification is a challenging research problem in soft biometrics and computer vision. In our work, we present a kinship verification system that starts with pair of facial images of the child and parent, then as a final result is determine whether two persons have a kin relation or not. our approach contains five steps as follows:(i) the face preprocessing step to get aligned and cropped facial images of the pair (ii), extracting deep features based on the deep learning model called Visual Geometry Group (VGG) Face,(iii) applying our proposed pair feature representation function alongside with a features normalization,(iv) the use of Fisher Score (FS) to select the best discriminative features,(v) decide whether there is a kinship or not based on the Support Vector Machine (SVM) classifier. We conducted several experiments to demonstrate the effectiveness of our approach that we …

BibTeX

@misc{CherguiNDKinshipVerification,
  title  = {Kinship Verification Through Facial Images Using CNN-Based Features Kinship Verification Through Facial Images Using CNN-Based Features},
  author = {Abdelhakim Chergui and Salim Ouchtati and Sébastien Mavromatis and Salah Eddine Bekhouche and Mohamed Lashab and Jean Sequeira},
  url    = {https://scholar.google.com/citations?view_op=view_citation&hl=en&user=IiMXAxUAAAAJ&pagesize=100&citation_for_view=IiMXAxUAAAAJ:_Qo2XoVZTnwC}
}