The safe and continuous operation of Unmanned Surface Vessels (USV) depends on robust visual perception systems capable of detecting and avoiding obstacles in a dynamic and unpredictable maritime environment. The main bottleneck for the development of such systems is the scarcity of comprehensive training datasets that capture the wide range of operational scenarios. This work proposes a programmatic methodology for generating large-scale synthetic datasets using a generative AI-based data augmentation approach. Our pipeline transforms a limited set of real images into a diverse dataset by introducing controlled variations in sea states, weather conditions, lighting, and a wide range of obstacles. The methodology includes modules to ensure image realism, validate the preservation of objects of interest, and mine "hard examples" to focus training on highly complex scenarios. We also propose an experimental design to validate the effectiveness of our generated dataset, comparing the performance of cutting-edge object detection architectures trained with raw data, data augmented by traditional methods, and data generated by our pipeline. The expected results indicate a significant improvement in the accuracy and robustness of the models trained with our method, paving the way for safer and more reliable anti-collision systems on autonomous robotic offshore platforms.

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.

Visual navigation for unmanned surface vehicles (USVs) remains challenging due to high-dimensionalvisual inputs, partial observability, and the difficulty of learning control-relevant representationswithout explicit scene reconstruction. Recent advances in joint-embedding predictive world modelssuggest that latent representations can capture task-relevant dynamics, but their use for reliableplanning and control is still not fully understood. This work proposes a theoretically groundedframework for latent-space planning using an action-conditioned joint-embedding predictive worldmodel. A video encoder maps short temporal clips into latent tokens, while an action-conditionedpredictor propagates these representations under candidate action sequences. Planning is formulatedas a discrete Model Predictive Control (MPC) problem, where motion primitives are selected from afinite library of interpretable trajectories through a composite cost function. The cost integrates goaldirectedbehavior, obstacle avoidance, latent predictive penalties, and stability constraints to ensuresmooth control execution. The central hypothesis is that if the learned latent representation capturestask-relevant invariances and control-sensitive factors, reliable planning can be performed directly inlatent space without explicit metric reconstruction. Experiments evaluate latent controllability, actionconditionedseparability, rollout fidelity across prediction horizons, and planning stability throughswitching frequency and control jitter metrics.

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

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. Code and Data Repository: Link

The integration of artificial intelligence (AI), particularly convolutional neural networks (CNNs), into dermatological diagnosis demonstrates substantial clinical potential. While the 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 whether 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 this subset, compared to 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 the misclassified subset. 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.

Anomaly detection in industrial contexts has become a critical task in computer vision due to its high relevance to industry applications. By grouping anomalies into coherent clusters based on their characteristics, which is commonly known as anomaly clustering, a better understanding of the anomalies and their causes is possible. We propose a novel anomaly clustering framework that leverages anomaly localization to create multi-scale views for contrastive learning. Unlike previous approaches, it integrates heatmap-guided multi-view feature extraction with self-supervised clustering, improving robustness and scalability in industrial defect analysis. Our approach utilizes spatial localization of anomalies by a state-of-the-art anomaly detector to generate multiple views of the image at different scales. These views are then used to train a convolutional neural network by means of contrastive learning, allowing the network to learn effective data representations. This technique enhances the network’s ability to distinguish between the unique features of anomalies from different perspectives, improving representation learning for anomaly clustering. We demonstrate the effectiveness of our method on benchmark datasets, showing an improvement in NMI of 12.5% in clustering accuracy for objects and textures on the MVTec-AD dataset. Our approach sets a new state-of-the-art, advancing the capabilities of anomaly clustering in industrial applications.

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.

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.

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/

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.

Purpose:Wireless capsule endoscopy (WCE) is a non-invasive technology used for diagnosing gastrointestinal abnormalities. A single examination generates ∼ 55,000 images, making manual review both time-consuming and costly for doctors. Therefore, the development of computer vision-assisted systems is highly desirable to aid in the diagnostic process. Approach:We presents a deep learning approach leveraging knowledge distillation (KD) from a convolutional neural network (CNN) teacher model to a vision transformer (ViT) student model for gastrointestinal abnormality recognition. The CNN teacher model utilizes attention mechanisms and depth-wise separable convolutions to extract features from WCE images, supervising the ViT in learning these representations. Results:The proposed method achieves accuracy of 97% and 96% on the Kvasir and KID datasets, respectively, demonstrating its effectiveness in distinguishing normal from abnormal regions and bleeding from non-bleeding cases. The proposed approach offers computational efficiency and generalization to unseen datasets, outperforming several state-of-the-art methods. Conclusions:We proposed a deep learning approach utilizing CNNs and a ViT with KD to effectively classify gastrointestinal diseases in WCE images. It demonstrates promising performance on public datasets, distinguishing normal from abnormal regions and bleeding from non-bleeding cases while offering optimal computational efficiency compared with existing methods, making it suitable for GI disease applications.

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.

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 better capture spatial dependencies and contextual information. 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.1 1 contains the data and code for all datasets, which are available for researchers and practitioners interested in replicating or building upon our work.

Segmentation of white blood cells is critical for a wide range of applications. It aims to identify and isolate individual white blood cells from medical images, enabling accurate diagnosis and monitoring of diseases. In the last decade, many researchers have focused on this task using U-Net, one of the most used deep learning architectures. To further enhance segmentation accuracy and robustness, recent advances have explored the combination of U-Net with other techniques, such as attention mechanisms and aggregation techniques. However, a common challenge in white blood cell image segmentation is the similarity between the cells’ cytoplasm and other surrounding blood components, which often leads to inaccurate or incomplete segmentation due to difficulties in distinguishing low-contrast or subtle boundaries, leaving a significant gap for improvement. In this paper, we propose GAAD-U-Net, a novel architecture that integrates attention-augmented convolutions to better capture ambiguous boundaries and complex structures such as overlapping cells and low-contrast regions, followed by a gating mechanism to further suppress irrelevant feature information. These two key components are integrated in the Double U-Net base architecture. Our model achieves state-of-the-art performance on white blood cell benchmark datasets, with a 3.4% Dice score coefficient (DSC) improvement specifically on the SegPC-2021 dataset. The proposed model achieves superior performance as measured by mean the intersection over union (IoU) and DSC, with notably strong segmentation performance even for difficult images.

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.

This review explores the recent advancements in AI-driven autonomous sailboat navigation, underscoring its pivotal role in ocean monitoring and real-time maritime data collection. Drawing on an extensive range of primary and secondary sources, the study critically evaluates current challenges, innovative control algorithms, and path planning strategies, with a particular emphasis on AI techniques. A major contribution of this study is the comparative analysis of these AI methods to assess their efficacy in achieving robust autonomy amid dynamic and uncertain maritime environments. The review also addresses notable gaps in the literature, highlighting the limited adoption of AI-specific methodologies in sailboat control systems. It explores hybrid and adaptive approaches that integrates advanced sensing and obstacle avoidance technologies to improve real-time decision-making and navigation accuracy. Furthermore, the paper traces the evolution of path planning from traditional graph-based methods to state-of-the-art learning algorithms, identifying future research directions focused on enhancing robustness, adaptability, and the practical deployment of autonomous sailboats beyond simulations. Ultimately, this review serves a foundational resource for researchers and practitioners aiming to advance sustainable, efficient, and reliable autonomous sailboat technologies for marine exploration and environmental Management.

The aging global population requires innovative remote monitoring systems to assist doctors and caregivers in assessing the health of elderly patients. Doctors often lack access to continuous behavioral data, making it difficult to detect deviations from normal patterns when elderly patients arrive for a consultation. Without historical insights into common behaviors and potential anomalies detected with unobtrusive techniques (e.g., non-wearable devices), timely and informed medical interventions become challenging. To address this, we propose an edge-based Internet of Things (IoT) framework that enables real-time monitoring and anomaly detection using non-wearable sensors to assist doctors and caregivers in assessing the health of elderly patients. By processing data locally, the system minimizes privacy concerns and ensures immediate data availability, allowing healthcare professionals to detect unusual behavioral patterns early. The system employs advanced machine learning (ML) models to identify deviations that may indicate potential health risks. A prototype of our system has been developed to test its feasibility and demonstrate, through the application of two of the most frequently used ML models, i.e., isolation forest and Long Short-Term Memory (LSTM) networks, that it can provide scalability, efficiency, and reliability in the context of elderly care. Further, the provided dashboard enables caregivers and healthcare professionals to access real-time alerts and longitudinal trends, facilitating proactive interventions. The proposed approach improves healthcare responsiveness by providing instant insights into patient behavior, facilitating more accurate diagnoses and interventions. This study lays the groundwork for future advancements in the field and offers valuable insights for the research community to harness the full potential of combining edge computing, artificial intelligence (AI), and the IoT in elderly care.