arXiv subjects

Cosimo Distante

Publications and source records attributed to Cosimo Distante.

At least 19 recordsLinked to original sources

arXiv2026

TMF-RSE: Tri-Modal Fusion with Regional Semantics and Evidential Uncertainty for Lung Severity Scoring

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.

eess.IV↗

arXiv2026

RF-HiT: Rectified Flow Hierarchical Transformer for General Medical Image Segmentation

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.

cs.CV↗

arXiv2026

When AI and Experts Agree on Error: Intrinsic Ambiguity in Dermatoscopic Images

The integration of artificial intelligence (AI), particularly Convolutional Neural Networks (CNNs), into dermatological diagnosis demonstrates substantial clinical potential. While existing literature predominantly benchmarks algorithmic performance against human experts, our study adopts a novel perspective by investigating the intrinsic complexity of dermatoscopic images. Through rigorous experimentation with multiple CNN architectures, we isolated a subset of images systematically misclassified across all models-a phenomenon statistically proven to exceed random chance. To determine if these failures stem from algorithmic biases or inherent visual ambiguity, expert dermatologists independently evaluated these challenging cases alongside a control group. The results revealed a collapse in human diagnostic performance on the AI-misclassified images. First, agreement with ground-truth labels plummeted, with Cohen's kappa dropping to a mere 0.08 for the difficult images, compared to a 0.61 for the control group. Second, we observed a severe deterioration in expert consensus; inter-rater reliability among physicians fell from moderate concordance (Fleiss kappa = 0.456) on control images to only modest agreement (Fleiss kappa = 0.275) on difficult cases. We identified image quality as a primary driver of these dual systematic failures. To promote transparency and reproducibility, all data, code, and trained models have been made publicly available

cs.CV↗

arXiv2026

Cross-Modal Mapping and Dual-Branch Reconstruction for 2D-3D Multimodal Industrial Anomaly Detection

Multimodal industrial anomaly detection benefits from integrating RGB appearance with 3D surface geometry, yet existing \emph{unsupervised} approaches commonly rely on memory banks, teacher-student architectures, or fragile fusion schemes, limiting robustness under noisy depth, weak texture, or missing modalities. This paper introduces \textbf{CMDR-IAD}, a lightweight and modality-flexible unsupervised framework for reliable anomaly detection in 2D+3D multimodal as well as single-modality (2D-only or 3D-only) settings. \textbf{CMDR-IAD} combines bidirectional 2D$\leftrightarrow$3D cross-modal mapping to model appearance-geometry consistency with dual-branch reconstruction that independently captures normal texture and geometric structure. A two-part fusion strategy integrates these cues: a reliability-gated mapping anomaly highlights spatially consistent texture-geometry discrepancies, while a confidence-weighted reconstruction anomaly adaptively balances appearance and geometric deviations, yielding stable and precise anomaly localization even in depth-sparse or low-texture regions. On the MVTec 3D-AD benchmark, CMDR-IAD achieves state-of-the-art performance while operating without memory banks, reaching 97.3\% image-level AUROC (I-AUROC), 99.6\% pixel-level AUROC (P-AUROC), and 97.6\% AUPRO. On a real-world polyurethane cutting dataset, the 3D-only variant attains 92.6\% I-AUROC and 92.5\% P-AUROC, demonstrating strong effectiveness under practical industrial conditions. These results highlight the framework's robustness, modality flexibility, and the effectiveness of the proposed fusion strategies for industrial visual inspection. Our source code is available at https://github.com/ECGAI-Research/CMDR-IAD/

cs.CV↗

arXiv2026

VP-Hype: A Hybrid Mamba-Transformer Framework with Visual-Textual Prompting for Hyperspectral Image Classification

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.

cs.CV↗

arXiv2025

VLM-PAR: A Vision Language Model for Pedestrian Attribute Recognition

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.

cs.CV↗

arXiv2025

T2IBias: Uncovering Societal Bias Encoded in the Latent Space of Text-to-Image Generative Models

Text-to-image (T2I) generative models are largely used in AI-powered real-world applications and value creation. However, their strategic deployment raises critical concerns for responsible AI management, particularly regarding the reproduction and amplification of race- and gender-related stereotypes that can undermine organizational ethics. In this work, we investigate whether such societal biases are systematically encoded within the pretrained latent spaces of state-of-the-art T2I models. We conduct an empirical study across the five most popular open-source models, using ten neutral, profession-related prompts to generate 100 images per profession, resulting in a dataset of 5,000 images evaluated by diverse human assessors representing different races and genders. We demonstrate that all five models encode and amplify pronounced societal skew: caregiving and nursing roles are consistently feminized, while high-status professions such as corporate CEO, politician, doctor, and lawyer are overwhelmingly represented by males and mostly White individuals. We further identify model-specific patterns, such as QWEN-Image's near-exclusive focus on East Asian outputs, Kandinsky's dominance of White individuals, and SDXL's comparatively broader but still biased distributions. These results provide critical insights for AI project managers and practitioners, enabling them to select equitable AI models and customized prompts that generate images in alignment with the principles of responsible AI. We conclude by discussing the risks of these biases and proposing actionable strategies for bias mitigation in building responsible GenAI systems. The code and Data Repository: https://github.com/Sufianlab/T2IBias

cs.LG↗

arXiv2025

CVPD at QIAS 2025 Shared Task: An Efficient Encoder-Based Approach for Islamic Inheritance Reasoning

Islamic inheritance law (Ilm al-Mawarith) 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 context-dependent. 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.

cs.CL↗

arXiv2025

C-DiffDet+: Fusing Global Scene Context with Generative Denoising for High-Fidelity Car Damage Detection

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

cs.CV↗

arXiv2025

DemoBias: An Empirical Study to Trace Demographic Biases in Vision Foundation Models

Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities across various downstream tasks, including biometric face recognition (FR) with description. However, demographic biases remain a critical concern in FR, as these foundation models often fail to perform equitably across diverse demographic groups, considering ethnicity/race, gender, and age. Therefore, through our work DemoBias, we conduct an empirical evaluation to investigate the extent of demographic biases in LVLMs for biometric FR with textual token generation tasks. We fine-tuned and evaluated three widely used pre-trained LVLMs: LLaVA, BLIP-2, and PaliGemma on our own generated demographic-balanced dataset. We utilize several evaluation metrics, like group-specific BERTScores and the Fairness Discrepancy Rate, to quantify and trace the performance disparities. The experimental results deliver compelling insights into the fairness and reliability of LVLMs across diverse demographic groups. Our empirical study uncovered demographic biases in LVLMs, with PaliGemma and LLaVA exhibiting higher disparities for Hispanic/Latino, Caucasian, and South Asian groups, whereas BLIP-2 demonstrated comparably consistent. Repository: https://github.com/Sufianlab/DemoBias.

cs.CV↗

arXiv2025

Beyond Linear Bottlenecks: Spline-Based Knowledge Distillation for Culturally Diverse Art Style Classification

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. Experiments on WikiArt and Pandora18k demonstrate that our approach outperforms the base dual teacher architecture in Top-1 accuracy. Our findings highlight the importance of KANs in disentangling complex style manifolds, leading to better linear probe accuracy than MLP projections.

cs.CV↗

arXiv2025

Markerless Stride Length estimation in Athletic using Pose Estimation with monocular vision

Performance measures such as stride length in athletics and the pace of runners can be estimated using different tricks such as measuring the number of steps divided by the running length or helping with markers printed on the track. Monitoring individual performance is essential for supporting staff coaches in establishing a proper training schedule for each athlete. The aim of this paper is to investigate a computer vision-based approach for estimating stride length and speed transition from video sequences and assessing video analysis processing among athletes. Using some well-known image processing methodologies such as probabilistic hough transform combined with a human pose detection algorithm, we estimate the leg joint position of runners. In this way, applying a homography transformation, we can estimate the runner stride length. Experiments on various race videos with three different runners demonstrated that the proposed system represents a useful tool for coaching and training. This suggests its potential value in measuring and monitoring the gait parameters of athletes.

cs.CV↗

arXiv2025

LoLA-SpecViT: Local Attention SwiGLU Vision Transformer with LoRA for Hyperspectral Imaging

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 \textbf{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 low-rank adaptation (LoRA) into attention and projection layers, enabling fine-tuning with over 80\% fewer trainable parameters. A novel cyclical learning rate scheduler modulates LoRA adaptation strength during training, improving convergence and generalisation. Extensive experiments on three benchmark datasets WHU-Hi LongKou, WHU-Hi HongHu, and Salinas 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. Our code is available in the following \href{https://github.com/FadiZidiDz/LoLA-SpecViT}{GitHub Repository}.

cs.CV↗

arXiv2025

Can Vision Transformers with ResNet's Global Features Fairly Authenticate Demographic Faces?

Biometric face authentication is crucial in computer vision, but ensuring fairness and generalization across demographic groups remains a big challenge. Therefore, we investigated whether Vision Transformer (ViT) and ResNet, leveraging pre-trained global features, can fairly authenticate different demographic faces while relying minimally on local features. In this investigation, we used three pre-trained state-of-the-art (SOTA) ViT foundation models from Facebook, Google, and Microsoft for global features as well as ResNet-18. We concatenated the features from ViT and ResNet, passed them through two fully connected layers, and trained on customized face image datasets to capture the local features. Then, we designed a novel few-shot prototype network with backbone features embedding. We also developed new demographic face image support and query datasets for this empirical study. The network's testing was conducted on this dataset in one-shot, three-shot, and five-shot scenarios to assess how performance improves as the size of the support set increases. We observed results across datasets with varying races/ethnicities, genders, and age groups. The Microsoft Swin Transformer backbone performed better among the three SOTA ViT for this task. The code and data are available at: https://github.com/Sufianlab/FairVitBio.

cs.CV↗

arXiv2025

Advancing Wheat Crop Analysis: A Survey of Deep Learning Approaches Using Hyperspectral Imaging

As one of the most widely cultivated and consumed crops, wheat is essential to global food security. However, wheat production is increasingly challenged by pests, diseases, climate change, and water scarcity, threatening yields. Traditional crop monitoring methods are labor-intensive and often ineffective for early issue detection. Hyperspectral imaging (HSI) has emerged as a non-destructive and efficient technology for remote crop health assessment. However, the high dimensionality of HSI data and limited availability of labeled samples present notable challenges. In recent years, deep learning has shown great promise in addressing these challenges due to its ability to extract and analysis complex structures. Despite advancements in applying deep learning methods to HSI data for wheat crop analysis, no comprehensive survey currently exists in this field. This review addresses this gap by summarizing benchmark datasets, tracking advancements in deep learning methods, and analyzing key applications such as variety classification, disease detection, and yield estimation. It also highlights the strengths, limitations, and future opportunities in leveraging deep learning methods for HSI-based wheat crop analysis. We have listed the current state-of-the-art papers and will continue tracking updating them in the following https://github.com/fadi-07/Awesome-Wheat-HSI-DeepLearning.

cs.CV↗

arXiv2025

Mamba Adaptive Anomaly Transformer with association discrepancy for time series

Anomaly detection in time series is essential for industrial monitoring and environmental sensing, yet distinguishing anomalies from complex patterns remains challenging. Existing methods like the Anomaly Transformer and DCdetector have progressed, but they face limitations such as sensitivity to short-term contexts and inefficiency in noisy, non-stationary environments. To overcome these issues, we introduce MAAT, an improved architecture that enhances association discrepancy modeling and reconstruction quality. MAAT features Sparse Attention, efficiently capturing long-range dependencies by focusing on relevant time steps, thereby reducing computational redundancy. Additionally, a Mamba-Selective State Space Model is incorporated into the reconstruction module, utilizing a skip connection and Gated Attention to improve anomaly localization and detection performance. Extensive experiments show that MAAT significantly outperforms previous methods, achieving better anomaly distinguishability and generalization across various time series applications, setting a new standard for unsupervised time series anomaly detection in real-world scenarios.

cs.LG↗

arXiv2024

Boosting Hyperspectral Image Classification with Gate-Shift-Fuse Mechanisms in a Novel CNN-Transformer Approach

During the process of classifying Hyperspectral Image (HSI), every pixel sample is categorized under a land-cover type. CNN-based techniques for HSI classification have notably advanced the field by their adept feature representation capabilities. However, acquiring deep features remains a challenge for these CNN-based methods. In contrast, transformer models are adept at extracting high-level semantic features, offering a complementary strength. This paper's main contribution is the introduction of an HSI classification model that includes two convolutional blocks, a Gate-Shift-Fuse (GSF) block and a transformer block. This model leverages the strengths of CNNs in local feature extraction and transformers in long-range context modelling. The GSF block is designed to strengthen the extraction of local and global spatial-spectral features. An effective attention mechanism module is also proposed to enhance the extraction of information from HSI cubes. The proposed method is evaluated on four well-known datasets (the Indian Pines, Pavia University, WHU-WHU-Hi-LongKou and WHU-Hi-HanChuan), demonstrating that the proposed framework achieves superior results compared to other models.

cs.CV↗

arXiv2024

Boosting House Price Estimations with Multi-Head Gated Attention

Evaluating house prices is crucial for various stakeholders, including homeowners, investors, and policymakers. However, traditional spatial interpolation methods have limitations in capturing the complex spatial relationships that affect property values. To address these challenges, we have developed a new method called Multi-Head Gated Attention for spatial interpolation. Our approach builds upon attention-based interpolation models and incorporates multiple attention heads and gating mechanisms to capture spatial dependencies and contextual information better. Importantly, our model produces embeddings that reduce the dimensionality of the data, enabling simpler models like linear regression to outperform complex ensembling models. We conducted extensive experiments to compare our model with baseline methods and the original attention-based interpolation model. The results show a significant improvement in the accuracy of house price predictions, validating the effectiveness of our approach. This research advances the field of spatial interpolation and provides a robust tool for more precise house price evaluation. Our GitHub repository.contains the data and code for all datasets, which are available for researchers and practitioners interested in replicating or building upon our work.

cs.LG↗