A5037064633- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Face and Expression Recognition
- Machine Learning and Data Classification
- Handwritten Text Recognition Techniques
- Video Analysis and Summarization
- Face recognition and analysis
- Biometric Identification and Security
- Natural Language Processing Techniques
- Image Processing and 3D Reconstruction
- Multimodal Machine Learning Applications
- COVID-19 diagnosis using AI
- AI in cancer detection
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Adversarial Robustness in Machine Learning
- Anomaly Detection Techniques and Applications
- Neural Networks and Applications
- Radiomics and Machine Learning in Medical Imaging
- Topic Modeling
- Meteorological Phenomena and Simulations
- Speech and Audio Processing
- Generative Adversarial Networks and Image Synthesis
- Domain Adaptation and Few-Shot Learning
- Biomedical Text Mining and Ontologies
Universitat Politècnica de València
2015-2024
Artificial Intelligence Research Institute
2023-2024
Centro Tecnológico de Investigación, Desarrollo e Innovación en tecnologías de la Información y las Comunicaciones (TIC)
2000-2011
Universitat de València
2010
In order to optimize the accuracy of nearest-neighbor classification rule, a weighted distance is proposed, along with algorithms automatically learn corresponding weights. These weights may be specific for each class and feature, individual prototype, or both. The learning are derived by (approximately) minimizing leaving-one-out error given training set. proposed approach assessed through series experiments UCI/STATLOG corpora, as well more task text which entails very sparse data...
Deep learning is revolutionizing radiology report generation (RRG) with the adoption of vision encoder–decoder (VED) frameworks, which transform radiographs into detailed medical reports. Traditional methods, however, often generate reports limited diversity and struggle generalization. Our research introduces reinforcement text augmentation to tackle these issues, significantly improving quality variability. By employing RadGraph as a reward metric innovating in augmentation, we surpass...
We present a new method for relevance feedback in image retrieval and scheme to learn weighted distances which can be used combination with different methods. User is crucial step maximise performance as was shown recent evaluations. Machine learning expected able how rank images according users needs. Most systems incorporate user using rather heuristic means only few groups have formally investigated the benefit from it machine techniques. our distance-learning into two approaches...
Person recognition using facial features, e.g., mug-shot images, has long been used in identity documents. However, due to the widespread use of web-cams and mobile devices embedded with a camera, it is now possible realize video recognition, rather than resorting just still images. In fact, offers many advantages over image recognition; these include potential boosting system accuracy deterring spoof attacks. This paper presents an evaluation person verification data, organized conjunction...
Given the overwhelming impact of machine learning on last decade, several libraries and frameworks have been developed in recent years to simplify design training neural networks, providing array-based programming, automatic differentiation user-friendly access hardware accelerators. None those tools, however, was designed with native transparent support for Cloud Computing or heterogeneous High-Performance (HPC). The DeepHealth Toolkit is an open source Deep Learning toolkit aimed at...
Pavement condition assessment is a critical step in road pavement management. In contrast to the automatic and objective methods used for rural roads, most commonly method urban areas development of visual surveys usually filled out by technicians, which leads subjective assessment. Whereas previous studies on identification distresses focused crack detection, this research aims not only cover classification multiple flexible (longitudinal transverse cracking, alligator raveling, potholes,...
This paper describes the ICFHR 2010 Contest for quantitative evaluation of binarization algorithms. These algorithm are applied to synthetic images modern pdf documents with noise from historical documents. Today, many scientists work on task and algorithms have been proposed. However, selection most appropriate one is not a simple procedure. The these proved be another difficult since there no objective way compare results. Here, 4 groups 6 systems participating in competition. experimental...
With this work we tackle immunofluorescence classification in renal biopsy, employing state-of-the-art Convolutional Neural Networks. In setting, the aim of probabilistic model is to assist an expert practitioner towards identifying location pattern antibody deposits within a glomerulus. Since modern neural networks often provide overconfident outputs, stress importance having reliable prediction, demonstrating that Temperature Scaling (TS), recently introduced re-calibration technique, can...
Abstract Convolutional neural networks (CNNs) have been broadly employed in dermoscopic image analysis, mainly as a result of the large amount data gathered by International Skin Imaging Collaboration (ISIC). As many other medical imaging domains, state‐of‐the‐art methods take advantage architectures developed for tasks, frequently assuming full transferability between enormous sets natural images (e.g. ImageNet) and images, which is not always case. A comprehensive analysis on effectiveness...