Cascade Transformers for End-to-End Person Search
This addresses the problem of accurately finding and identifying people in images for applications like surveillance or security, representing an incremental improvement over existing methods.
The paper tackles the problem of person search, which involves localizing a target person from scene images despite challenges like scale variations and occlusions, by proposing the Cascade Occluded Attention Transformer (COAT) that achieves state-of-the-art performance on two benchmark datasets.
The goal of person search is to localize a target person from a gallery set of scene images, which is extremely challenging due to large scale variations, pose/viewpoint changes, and occlusions. In this paper, we propose the Cascade Occluded Attention Transformer (COAT) for end-to-end person search. Our three-stage cascade design focuses on detecting people in the first stage, while later stages simultaneously and progressively refine the representation for person detection and re-identification. At each stage the occluded attention transformer applies tighter intersection over union thresholds, forcing the network to learn coarse-to-fine pose/scale invariant features. Meanwhile, we calculate each detection's occluded attention to differentiate a person's tokens from other people or the background. In this way, we simulate the effect of other objects occluding a person of interest at the token-level. Through comprehensive experiments, we demonstrate the benefits of our method by achieving state-of-the-art performance on two benchmark datasets.