Abstract
In this paper, we propose combined visual features for person re-identification. Our features are based on the multiple hand-crafted visual features. The proposed features are a combination of histogram from the RGB, YUV and HSV color channels, LBP and SIFT features. Then we use different distance metric learning methods to measure the similarity of the same persons and different persons. Experimental results demonstrate that the combined features have discriminative power for person re-identification.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have