Appearance-Preserving 3D Convolution for Video-based Person Re-identification
Preprint
Abstract
Due to the imperfect person detection results and posture changes, temporal appearance misalignment is unavoidable in video-based person re-identification (ReID). In this case, 3D convolution may destroy the appearance representation of person video clips, thus it is harmful to ReID. To address this problem, we propose AppearancePreserving 3D Convolution (AP3D), which is composed of two components: an Appearance-Preserving Module (APM) and a 3D convolution kernel. With APM aligning the adjacent feature maps in pixel level, the following 3D convolution can model temporal information on the premise of maintaining the appearance representation quality. It is easy to combine AP3D with existing 3D ConvNets by simply replacing the original 3D convolution kernels with AP3Ds. Extensive experiments demonstrate the effectiveness of AP3D for video-based ReID and the results on three widely used datasets surpass the state-of-the-arts. Code is available at: https://github.com/guxinqian/AP3D.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Temporal Complementary Learning for Video Person Re-identification2020
- Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training2020
- Video-Based Person Re-identification via 3D Convolutional Networks and Non-local Attention2019
- Beyond Part Models: Person Retrieval with Refined Part Pooling (and A Strong Convolutional Baseline)2018
- Quality Aware Network for Set to Set Recognition2017
- Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset2017
- Deep Residual Learning for Image Recognition2016
- Recurrent Convolutional Network for Video-Based Person Re-identification2016
- Beyond short snippets: Deep networks for video classification2015
- Person Re-identification by Video Ranking2014