The textile industry faces economic and environmental challenges due to low recycling rates and contamination from fasteners like buttons, rivets, and zippers. This paper proposes an Red, Green, Blue (RGB) vision system using You Only Look Once version 11 (YOLOv11) with a sliding window technique for automated fastener detection. The system addresses small object detection, occlusion, and fabric variability, incorporating Grounding DINO for garment localization and U2-Net for segmentation. Experiments show the sliding window method outperforms full-image detection for buttons and rivets (precision 0.874, recall 0.923), while zipper detection is less effective due to dataset limitations. This work advances scalable AI-driven solutions for textile recycling, supporting circular economy goals. Future work will target hidden fasteners, dataset expansion and fastener removal.

Abstract Radiographic canine hip dysplasia (CHD) diagnosis is crucial for breeding selection and disease management, delaying progression and alleviating the associated pain. Radiography is the primary imaging modality for CHD diagnosis, and visual assessment of radiographic features is sometimes used for accurate diagnosis. Specifically, alterations in femoral neck shape are crucial radiographic signs, with existing literature suggesting that dysplastic hips have a greater femoral neck thickness (FNT). In this study we aimed to develop a three-stage deep learning-based system that can automatically identify and quantify a femoral neck thickness index (FNTi) as a key metric to improve CHD diagnosis. Our system trained a keypoint detection model and a segmentation model to determine landmark and boundary coordinates of the femur and acetabulum, respectively. We then executed a series of mathematical operations to calculate the FNTi. The keypoint detection model achieved a mean absolute error (MAE) of 0.013 during training, while the femur segmentation results achieved a dice score (DS) of 0.978. Our three-stage deep learning-based system achieved an intraclass correlation coefficient of 0.86 (95% confidence interval) and showed no significant differences in paired t-test compared to a specialist (p > 0.05). As far as we know, this is the initial study to thoroughly measure FNTi by applying computer vision and deep learning-based approaches, which can provide reliable support in CHD diagnosis.

Coronary artery disease (CAD) remains a leading cause of morbidity and mortality worldwide, underscoring the need for accurate and reliable diagnostic tools. While AI-driven models have shown significant promise in identifying CAD through imaging techniques, their 'black box' nature often hinders clinical adoption due to a lack of interpretability. In response, this paper proposes a novel approach to image captioning specifically tailored for CAD diagnosis, aimed at enhancing the transparency and usability of AI systems. Utilizing the COCA dataset, which comprises gated coronary CT images along with Ground Truth (GT) segmentation annotations, we introduce a hybrid model architecture that combines a Vision Transformer (ViT) for feature extraction with a Generative Pretrained Transformer (GPT) for generating clinically relevant textual descriptions. This work builds on a previously developed 3D Convolutional Neural Network (CNN) for coronary artery segmentation, leveraging its accurate delineations of calcified regions as critical inputs to the captioning process. By incorporating these segmentation outputs, our approach not only focuses on accurately identifying and describing calcified regions within the coronary arteries but also ensures that the generated captions are clinically meaningful and reflective of key diagnostic features such as location, severity, and artery involvement. This methodology provides medical practitioners with clear, context-rich explanations of AI-generated findings, thereby bridging the gap between advanced AI technologies and practical clinical applications. Furthermore, our work underscores the critical role of Explainable AI (XAI) in fostering trust, improving decision-making, and enhancing the efficacy of AI-driven diagnostics, paving the way for future advancements in the field.

Public lighting plays a very important role for society's safety and quality of life. The identification of faults in public lighting is essential for the maintenance and prevention of safety. Traditionally, this task depends on human action, through checking during the day, representing expenditure and waste of energy. Automatic detection with deep learning is an innovative solution that can be explored for locating and identifying of this kind of problem. In this study, we present a first approach, composed of several steps, intending to obtain the segmentation of public lighting, using Seville (Spain) as case study. A dataset called NLight was created from a nighttime image taken by the JL1-3B satellite, and four U-Net and FPN architectures were trained with different backbones to segment part of the NLight. The U-Net with InceptionResNetv2 proved to be the model with the best performance, obtained 761 of 815, correct locations (93.4%). This model was used to predict the segmentation of the remaining dataset. This study provides the location of lamps so that we can identify patterns and possible lighting failures in the future.

This paper presents an innovative and motivating methodology to learn vision systems using a humanoid robot, NAO robot. Vision systems are an area of growing development and interest of engineering students. This approach to learning was applied in students of Master of Electrical Engineering. The goal is to introduce students the main approaches of visual object recognition and human face recognition using computer vision techniques to be embedded in a social robot and therefore he is able to interact with human beings. NAO robot as an educational platform easy to learn how to program, and it has a high sensory ability and two cameras that can capture the images for processing.

In this work, image segmentation is addressed as the starting point within a motion analysis methodology intended for biomechanics behavior characterization. First, we propose a general segmentation framework that uses Atanassov's intuitionistic fuzzy sets (A-IFSs) to determine the optimal image threshold value. Atanassov's intuitionistic fuzzy index values are used for representing the unknowledge/ignorance of an expert on determining whether a pixel belongs to the background or the object of the image. Then, we introduce an extension of this methodology that uses a heuristic based multi-threshold approach to determine the optimal threshold. Experimental results are presented.