Semantic Map Guided Identity Transfer GAN for Person Re-identification
Open Access
- 12 September 2024
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 20 (11) , 1-20
Abstract
Generative adversarial networks (GANs)-based person re-identification (re-id) schemes provide potential ways to augment data in practical applications. However, existing solutions perform poorly because of the separation of data generation and re-id training and a lack of diverse data in real-world scenarios. In this paper, a person re-id model (IDGAN) based on semantic map guided identity transfer GAN is proposed to improve the person re-id performance. With the aid of the semantic map, IDGAN generates pedestrian images with varying poses, perspectives, and backgrounds efficiently and accurately, improving the diversity of training data. To increase the visual realism, IDGAN utilizes a gradient augmentation method based on local quality attention to refine the generated image locally. Then, a two-stage joint training framework is employed to allow the GAN and the person re-id network to learn from each other to better use the generated data. Detailed experimental results demonstrate that, compared with the existing state-of-the-art methods, IDGAN is capable of producing high-quality images and significantly enhancing re-id performance, with the FID of generated images on the Market-1501 dataset being reduced by 1.15, and mAP on the Market-1501 and DukeMTMC-reID datasets being increased by 3.3% and 2.6%, respectively.Keywords
Funding Information
- National Key R&D Program of China (2020YFC1522900)
- HZAU-AGIS Cooperation Fund (SZYJY2023011)
- Knowledge Innovation Program of Wuhan-Basic Research (2022010801010225)
- Fundamental Research Funds
- Central Universities (2662022YJ012)
- National Natural Science Foundation of China (No. 61772562 and No. 62172438)
- Key Project of Shenzhen City Special Fund for Fundamental Research (JCYJ20220818103200002)
This publication has 35 references indexed in Scilit:
- Deep Fusion Feature Representation Learning With Hard Mining Center-Triplet Loss for Person Re-IdentificationIEEE Transactions on Multimedia, 2020
- Deep-Person: Learning discriminative deep features for person Re-IdentificationPattern Recognition, 2020
- Attributes-aided part detection and refinement for person re-identificationPattern Recognition, 2020
- Multi-Pseudo Regularized Label for Generated Data in Person Re-IdentificationIEEE Transactions on Image Processing, 2018
- The Unreasonable Effectiveness of Deep Features as a Perceptual Metric2018
- Look into Person: Joint Body Parsing & Pose Estimation Network and a New BenchmarkIEEE Transactions on Pattern Analysis and Machine Intelligence, 2018
- Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in Vitro2017
- Fully Deep Blind Image Quality PredictorIEEE Journal of Selected Topics in Signal Processing, 2016
- ImageNet: A large-scale hierarchical image database2009
- Image quality assessment: from error visibility to structural similarityIEEE Transactions on Image Processing, 2004