Color inference from semantic labeling for person search in videos

11/29/2019
by   Jules Simon, et al.
0

We propose an explainable model to generate semantic color labels for person search. In this context, persons are described from their semantic parts, such as hat, shirt, etc. Person search consists in looking for people based on these descriptions. In this work, we aim to improve the accuracy of color labels for people. Our goal is to handle the high variability of human perception. Existing solutions are based on hand-crafted features or learnt features that are not explainable. Moreover most of them only focus on a limited set of colors. We propose a method based on binary search trees and a large peer-labelled color name dataset. This allows us to synthesize the human perception of colors. Using semantic segmentation and our color labeling method, we label segments of pedestrians with their associated colors. We evaluate our solution on person search on datasets such as PCN, and show a precision as high as 80.4

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset