A5004665770- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Medical Image Segmentation Techniques
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Retinal Imaging and Analysis
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Image Processing Techniques and Applications
- Video Surveillance and Tracking Methods
- Human Pose and Action Recognition
- Digital Imaging for Blood Diseases
- Nutritional Studies and Diet
- Retinal Diseases and Treatments
- Anomaly Detection Techniques and Applications
- Infrared Thermography in Medicine
- Optical measurement and interference techniques
- Image and Signal Denoising Methods
- Advanced Image Processing Techniques
- COVID-19 diagnosis using AI
- Industrial Vision Systems and Defect Detection
- Brain Tumor Detection and Classification
- Cutaneous Melanoma Detection and Management
- Robotic Locomotion and Control
- Robotics and Sensor-Based Localization
Universitat Rovira i Virgili
2016-2025
Deakin University
2022
Université de Bourgogne
2022
Hospital Sant Joan de Déu Barcelona
2021
International Crisis Group
2019
Electrodermal activity (EDA) is indicative of psychological processes related to human cognition and emotions. Previous research has studied many methods for extracting EDA features; however, their appropriateness emotion recognition been tested using a small number distinct feature sets on different, usually small, data sets. In the current research, we reviewed 25 studies implemented 40 different features across time, frequency time-frequency domains publicly available AMIGOS dataset. We...
The limited depth-of-field of some cameras prevents them from capturing perfectly focused images when the imaged scene covers a large distance range. In order to compensate for this problem, image fusion has been exploited combining captured with different camera settings, thus yielding higher quality all-in-focus image. Since most current approaches rely on maximizing spatial frequency composed image, process is sensitive noise. paper, new algorithm computing sequence low presented....
Breast density estimation with visual evaluation is still challenging due to low contrast and significant fluctuations in the mammograms' fatty tissue background. The primary key breast classification detect dense tissues mammographic images correctly. Many methods have been proposed for estimation; nevertheless, most of them are not fully automated. Besides, they badly affected by signal-to-noise ratio variability appearance texture. This study intends develop a automated digitalized...
To solve real-life problems for different smart city applications, using deep Neural Network, such as parking occupancy detection, requires fine-tuning of these networks. For large parking, it is desirable to use a cenital-plane camera located at high distance that allows the monitoring entire space or area with only one camera. Today's most popular object detection models, YOLO, achieve good precision scores real-time speed. However, if we our own data from general-purpose datasets, COCO...
Depth estimation is a challenging task of 3D reconstruction to enhance the accuracy sensing environment awareness. This work brings new solution with improvements, which increases quantitative and qualitative understanding depth maps compared existing methods. Recently, convolutional neural networks (CNN) have demonstrated their extraordinary ability estimate from monocular videos. However, traditional CNN does not support topological structure, they can only on regular image regions...
Breast cancer needs to be detected early reduce mortality rate. Ultrasound imaging (US) could significantly enhance diagnosing cases with dense breasts. Most of the existing computer-aided diagnosis (CAD) systems employ a single ultrasound image for breast tumor extract features classify it as benign or malignant. However, accuracy such CAD system is limited due large size and shape variation, irregular ambiguous boundaries, low signal-to-noise ratio in images their noisy nature significant...
In-utero fetal MRI is emerging as an important tool in the diagnosis and analysis of developing human brain. Automatic segmentation brain a vital step quantitative prenatal neurodevelopment both research clinical context. However, manual cerebral structures time-consuming prone to error inter-observer variability. Therefore, we organized Fetal Tissue Annotation (FeTA) Challenge 2021 order encourage development automatic algorithms on international level. The challenge utilized FeTA Dataset,...
In recent years, several deep learning models have been proposed to accurately quantify and diagnose cardiac pathologies. These automated tools heavily rely on the accurate segmentation of structures in MRI images. However, right ventricle is challenging due its highly complex shape ill-defined borders. Hence, there a need for new methods handle such structure's geometrical textural complexities, notably presence pathologies as Dilated Right Ventricle, Tricuspid Regurgitation,...
Nowadays, breast cancer is one of the most common cancers diagnosed in women. Mammography standard screening imaging technique for early detection cancer. However, thermal infrared images (thermographies) can be used to reveal lesions dense breasts. In these images, temperature regions that contain tumors warmer than normal tissue. To detect difference between and cancerous regions, a dynamic thermography procedure uses cameras generate at fixed time steps, obtaining sequence images. this...
Skin lesion segmentation in dermoscopic images is still a challenge due to the low contrast and fuzzy boundaries of lesions. Moreover, lesions have high similarity with healthy regions terms appearance. In this paper, we propose an accurate skin model based on modified conditional generative adversarial network (cGAN). We introduce new block encoder cGAN called factorized channel attention (FCA), which exploits both mechanism residual 1-D kernel convolution. The increases discriminability...
The determination of precise skin lesion boundaries in dermoscopic images using automated methods faces many challenges, most importantly, the presence hair, inconspicuous edges and low contrast images, variability color, texture shapes lesions. Existing deep learning-based segmentation algorithms are expensive terms computational time memory. Consequently, running such requires a powerful GPU high bandwidth memory, which not available dermoscopy devices. Thus, this article aims to achieve...
Most of the variational optical flow methods are based on well-known brightness constancy assumption or high-order assumptions to implement data term in optimization energy function. Unfortunately, any variation lighting within scene violates constraint; turn, gradient does not work properly with large illumination changes. This paper proposes an illumination-robust a robust texture descriptor rather than constancy. Thus, similarity function used as was obtained from extracting features...
Skin cancer is the most common tumor in population. There are different therapeutic modalities. Brachytherapy one of techniques used, which it necessary to build customized moulds for some patients. Currently, these made by hand using rudimentary techniques. We present a new procedure based on 3D printing and analysis clinical workflow.Moulds can be either or automated printing. For making hand, patient's alginate negative created and, from that, gypsum cast template. The process first step...