Visual Understanding of Outdoor Dynamic Scenes


Video captured freely in outdoor environments pose significant challenges for automated visual understanding; that is, the recognition and localisation of semantic objects by a computer.

The work here aims to advance state of the art; the key-challenge is to segment video into identifiable components that are of relevance to the other sub-projects within OAK. This unit is led by Dr Matthew Brown (University of Bath) and aims at identifying and localising objects of interest within dynamic scenes.