

Development of predictive artificial intelligence models based on machine learning algorithms
The CVBLab research group is an expert in the development of machine learning algorithms and, more specifically, deep learning. CVBLab uses deep learning techniques to build prediction models based on examples such as simulating human behaviour in terms of neural connections. The development of a predictive model consists of training algorithms (frequently, neural networks) to solve a specific task (object detection, classification, segmentation, reconstruction, pattern recognition, etc.) using certain input data. Training can be carried out following different learning methodologies depending on the amount of labelled data available. CVBLab bases its experience on the research, development, and implementation of machine learning algorithms applicable in different areas such as health, industry, or agriculture, among others.
Applications:
• Diagnostic aid systems to automatically predict the severity of diseases of various kinds such as prostate, breast, skin, bladder, colon cancer, etc., diabetes, glaucoma, and diabetic retinopathy, among others. To do this, CVBLab works with different types of digital imaging such as histological imaging, hyperspectral imaging, infrared imaging, fundus, or optical coherence tomography.
• Prediction systems to analyse the feasibility of embryo implantation during the in vitro fertilization procedure.
• Real-time monitoring systems aimed at tracking tasks, stock control using drones, digital assistants for airplane cabins, people detection to measure social distance or object detection to control traffic and measure pollution levels, among other applications.
• Analysis, processing and processing of signals, images, and videos (1D, 2D and 3D data).
• Analysis systems applied to the agri-food sector, such as intelligent crop control using ortho-imaging, prediction of optimal ripeness of fruits and vegetables from hyperspectral imaging, and detection of artifacts in food packaging.
• Pattern recognition from large amounts of structured data, applying big data techniques.
• Artificial intelligence systems aimed at the “”business and analytics”” sector for stock market prediction, estimation of the price of a house, recommendation of personalized products to the customer’s profile, etc.
• Intelligent zero-defect systems dedicated to the industrial sector to be used in various manufacturing processes such as mass production or batch manufacturing.