Transferring a semantic representation for person re-identification and search

Learning semantic attributes for person re-identification and description-based person search has gained increas-ing interest due to attributes ’ great potential as a pose and view-invariant representation. However, existing attribute-centric approaches have thus far underperformed state-of-the-art conventional approaches. This is due to their non-scalable need for extensive domain (camera) specific an-notation. In this paper we present a new semantic at-tribute learning approach for person re-identification and search. Our model is trained on existing fashion photog-raphy datasets – either weakly or strongly labelled. It can then be transferred and adapted to provide a pow-erful semantic description of surveillance person detec-tions, without requiring any surveillance domain supervi-sion. The resulting representation is useful for both un-supervised and supervised person re-identification, achiev-ing state-of-the-art and near state-of-the-art performance respectively. Furthermore, as a semantic representation it allows description-based person search to be integrated within the same framework. 1