Yulia Brand, (EE, Technion)
Tuesday, 13.5.2014, 11:30
The person re-identification (ReID) problem has lately received increasing attention especially due to its important role in surveillance systems, which should be able to keep track of people after they have left the field of view of one camera and entered the field of view of any overlapping or non-overlapping camera.
It was shown that person ReID accuracy can be significantly improved given a training set that demonstrates changes in appearances associated with the two non-overlapping cameras involved. Here we test whether this advantage can be maintained when directly annotated training sets are not available for all camera-pairs at the site. Given the training set capturing correspondences between cameras A and B and a different training set capturing correspondences between cameras B and C, we suggest the Transitive Re-Identification algorithm (TRID) that provides a classifier for (A,C) appearance pairs. The proposed method is based on statistical modeling and uses a marginalization process for the inference. This approach significantly reduces the annotation effort inherent in a learning system, which goes down from O(N^2) to O(N), for a site containing N cameras. Moreover, when adding camera (N+1), only one inter-camera training set is required for establishing all correspondences. In our experiments we found
that the method is effective and more accurate than the competing camera invariant approach.
MSc research under the supervision of Prof. Michael Lindenbaum and Dr. Tamar Avraha