Multidirection and Multiscale Pyramid in Transformer for Video-Based Pedestrian Retrieval

Abstract

In video surveillance, pedestrian retrieval (also called person reidentification) is a critical task. This task aims to retrieve the pedestrian of interest from nonoverlapping cameras. Recently, transformer-based models have achieved significant progress for this task. However, these models still suffer from ignoring fine-grained, part-informed information. This article proposes a multidirection and multiscale Pyramid in Transformer (PiT) to solve this problem. In transformer-based architecture, each pedestrian image is split into many patches. Then, these patches are fed to transformer layers to obtain the feature representation of this image. To explore the fine-grained information, this article proposes to apply vertical division and horizontal division on these patches to generate different-direction human parts. These parts provide more fine-grained information. To fuse multiscale feature representation, this article presents a pyramid structure containing global-level information and many pieces of local-level information from different scales. The feature pyramids of all the pedestrian images from the same video are fused to form the final multidirection and multiscale feature representation. Experimental results on two challenging video-based benchmarks, MARS and iLIDS-VID, show the proposed PiT achieves state-of-the-art performance. Extensive ablation studies demonstrate the superiority of the proposed pyramid structure. Data is available on-line at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://git.openi.org.cn/zangxh/PiT.git</uri> .