During almost 15 years, the GVIS (formerly VARP) research group has worked on methods related to Computer Vision, Machine Learning and Deep Learning with the goal of solving multidisciplinary problems. Recently, we have started to work also on Natural Language Processing.

The problems we have dealt with can be grouped into three main topics, which constitutes our research lines: Cybersecurity and Crime prevention, medical and veterinary image analysis, and quality control in industry.

We have worked on applying Computer Vision and Machine Learning to biomedical image analysis problems: collaborating with veterinary experts for assessing automatically boar semen quality for artificial insemination purposes, developing systems to automatically detect droplets and whiptails in boar sperm by means of automatically identifying healthy tissues in the cardiovascular system.

In collaboration with a team of Mechanical Engineers, we have applied Computer Vision and Machine Learning techniques for successfully assessing surface roughness automatically without using any contact device, or estimating life of cutting tools in milling processes with no need of human intervention, to predict the best time to replace a worn tool. Currently, we are trying to apply it to assess the surface quality of tools built using additive manufacturing with metallic and ceramic materials.

In the context of cybersecurity and crime prevention, we are collaborating with the Spanish Institute of Cybersecurity (INCIBE) to research and develop tools based on Computer Vision, Natural Language Processing, Machine Learning and Deep Learning to fight against cybercrime (e.g. age estimation methods, object detection and recognition, classification of website visual and textual contents, perceptual hashing for darknet services identification), and for cybersecurity purposes (e.g. botnet detection).