Date: Wednesday, June 4th, 2014
Place: Informatics Teaching Lab (ITL) — Top floor seminar room
Speaker: Ryan Layne
Title: Domain Transfer for Person Re-identification
Automatic person re-identification is a crucial capability under-pinning many applications in public video surveillance. It is challenging due to intra-class variation in person appearance when observed in different views, together with limited inter-class variability. Various recent approaches have made great progress in re-identification performance using discriminative techniques. However, these approaches are fundamentally limited by the requirement of extensive annotated training data for every pair of views. For practical re-identification. this is an unreasonable assumption, as annotating extensive volumes of data for every pair of cameras to be re-identified is often impossible or prohibitively expensive, and some specific applications for re-identification may deny the opportunity for any extensive training at all.
In this paper we move toward relaxing this strong assumption by investigating flexible multi-source transfer of re-identification models across pairs specifically, we show how to leverage prior re-identification models learned for a set of source view pairs (domains), and flexibly combine these to obtain good re-identification performance in a target view pair (domain) with greatly reduced training data requirements in the target domain.
Ryan Layne is a 3rd year PhD student in our group.