Ordered or Orderless: A Revisit for Video based Person Re-Identification

TL;DR

This paper challenges the necessity of recurrent networks in video person re-identification, proposing an orderless ensemble approach that treats videos as collections of images, leading to state-of-the-art results.

Contribution

It introduces a simple, effective orderless ensemble method for VPRe-id, moving away from recurrent structures and bridging video and image re-identification.

Findings

Recurrent networks may not effectively learn temporal dependencies in VPRe-id.

Treating videos as image collections improves re-identification performance.

The proposed method achieves state-of-the-art results on multiple datasets.

Abstract

Is recurrent network really necessary for learning a good visual representation for video based person re-identification (VPRe-id)? In this paper, we first show that the common practice of employing recurrent neural networks (RNNs) to aggregate temporal spatial features may not be optimal. Specifically, with a diagnostic analysis, we show that the recurrent structure may not be effective to learn temporal dependencies than what we expected and implicitly yields an orderless representation. Based on this observation, we then present a simple yet surprisingly powerful approach for VPRe-id, where we treat VPRe-id as an efficient orderless ensemble of image based person re-identification problem. More specifically, we divide videos into individual images and re-identify person with ensemble of image based rankers. Under the i.i.d. assumption, we provide an error bound that sheds light upon how could we improve VPRe-id. Our work also presents a promising way to bridge the gap between video and image based person re-identification. Comprehensive experimental evaluations demonstrate that the proposed solution achieves state-of-the-art performances on multiple widely used datasets (iLIDS-VID, PRID 2011, and MARS).