The scientific objective of the Image & Vision area is to develop new conceptual approaches and adapt theoretical tools to solve complex problems in digital image processing and computer vision.
Research themes
- Multimodal Image Analysis: This theme covers the Image & Vision area’s work on the processing and analysis of images from different modalities (visible, multispectral, hyperspectral, thermal, ultrasound, microscopic, 3D, etc.). These types of imaging cover the axis' applications, which are mainly governed by research projects. On the scientific side, these include multimodal image registration, 3D point processing, feature extraction, segmentation, classification, detection, recognition, tracking, behavioral analysis, fusion, etc. The scientific bottlenecks are accentuated by these specific imaging modalities, where conventional algorithms and approaches often fail to achieve the desired ends.
- Machine Learning & Vision: The Image & Vision area works actively with machine learning tools, in particular deep neural networks for analyzing images from different modalities. The aim is to develop approaches, concepts, methodologies and architectures to address the various problems associated with deep learning.