The University of Southampton
Computer Vision

VLC’s research in image processing and computer vision spans techniques from preprocessing, to feature extraction and on to image analysis. Approaches to feature extraction have extended classic technique, such as active contours/ level sets and the Hough transform, and have started totally new approaches using analogies such as water, heat and light. VLC researchers have a long record in biometrics, pioneering work in gait and facial recognition. Work continues in these areas and VLC are now developing soft biometrics, learning from human labelling to augment or even replace the automatically derived measures.

Machine Learning

Machine learning in VLC covers a broad range of areas ranging from developing new classification and clustering tools for big data sets, mathematical modelling of complex systems and optimisation.  Application areas include biological sequence analysis, gene regulation, text analysis, recommender systems and combinatorial optimisation.

Control

VLC researchers’ work on fundamental theory includes behavioural approaches to system theory, system identification particularly using structured low-rank approximations, multidimensional systems theory, robust nonlinear control, iterative learning control, adaptive control and flow control. They use a variety of techniques from linear algebra, functional analysis, partial differential equations and commutative algebra.