IEICE TRANSACTIONS on Information
Multi-Scale Correspondence Learning for Person Image Generation
Summary :
A generative model is presented for two types of person image generation in this paper. First, this model is applied to pose-guided person image generation, i.e., converting the pose of a source person image to the target pose while preserving the texture of that source person image. Second, this model is also used for clothing-guided person image generation, i.e., changing the clothing texture of a source person image to the desired clothing texture. The core idea of the proposed model is to establish the multi-scale correspondence, which can effectively address the misalignment introduced by transferring pose, thereby preserving richer information on appearance. Specifically, the proposed model consists of two stages: 1) It first generates the target semantic map imposed on the target pose to provide more accurate guidance during the generation process. 2) After obtaining the multi-scale feature map by the encoder, the multi-scale correspondence is established, which is useful for a fine-grained generation. Experimental results show the proposed method is superior to state-of-the-art methods in pose-guided person image generation and show its effectiveness in clothing-guided person image generation.
- Publication
- IEICE TRANSACTIONS on Information Vol.E106-D No.5 pp.804-812
- Publication Date
- 2023/05/01
- Publicized
- 2022/04/15
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.2022DLP0058
- Type of Manuscript
- Special Section PAPER (Special Section on Deep Learning Technologies: Architecture, Optimization, Techniques, and Applications)
- Category
- Person Image Generation
Authors
Shi-Long SHEN
Harbin Institute of Technology (Shenzhen)
Ai-Guo WU
Harbin Institute of Technology (Shenzhen)
Yong XU
Shenzhen Key Laboratory of Visual Object Detection and Recognition
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Shi-Long SHEN, Ai-Guo WU, Yong XU, "Multi-Scale Correspondence Learning for Person Image Generation" in IEICE TRANSACTIONS on Information,
vol. E106-D, no. 5, pp. 804-812, May 2023, doi: 10.1587/transinf.2022DLP0058.
Abstract: A generative model is presented for two types of person image generation in this paper. First, this model is applied to pose-guided person image generation, i.e., converting the pose of a source person image to the target pose while preserving the texture of that source person image. Second, this model is also used for clothing-guided person image generation, i.e., changing the clothing texture of a source person image to the desired clothing texture. The core idea of the proposed model is to establish the multi-scale correspondence, which can effectively address the misalignment introduced by transferring pose, thereby preserving richer information on appearance. Specifically, the proposed model consists of two stages: 1) It first generates the target semantic map imposed on the target pose to provide more accurate guidance during the generation process. 2) After obtaining the multi-scale feature map by the encoder, the multi-scale correspondence is established, which is useful for a fine-grained generation. Experimental results show the proposed method is superior to state-of-the-art methods in pose-guided person image generation and show its effectiveness in clothing-guided person image generation.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2022DLP0058/_p
Copy
@ARTICLE{e106-d_5_804,
author={Shi-Long SHEN, Ai-Guo WU, Yong XU, },
journal={IEICE TRANSACTIONS on Information},
title={Multi-Scale Correspondence Learning for Person Image Generation},
year={2023},
volume={E106-D},
number={5},
pages={804-812},
abstract={A generative model is presented for two types of person image generation in this paper. First, this model is applied to pose-guided person image generation, i.e., converting the pose of a source person image to the target pose while preserving the texture of that source person image. Second, this model is also used for clothing-guided person image generation, i.e., changing the clothing texture of a source person image to the desired clothing texture. The core idea of the proposed model is to establish the multi-scale correspondence, which can effectively address the misalignment introduced by transferring pose, thereby preserving richer information on appearance. Specifically, the proposed model consists of two stages: 1) It first generates the target semantic map imposed on the target pose to provide more accurate guidance during the generation process. 2) After obtaining the multi-scale feature map by the encoder, the multi-scale correspondence is established, which is useful for a fine-grained generation. Experimental results show the proposed method is superior to state-of-the-art methods in pose-guided person image generation and show its effectiveness in clothing-guided person image generation.},
keywords={},
doi={10.1587/transinf.2022DLP0058},
ISSN={1745-1361},
month={May},}
Copy
TY - JOUR
TI - Multi-Scale Correspondence Learning for Person Image Generation
T2 - IEICE TRANSACTIONS on Information
SP - 804
EP - 812
AU - Shi-Long SHEN
AU - Ai-Guo WU
AU - Yong XU
PY - 2023
DO - 10.1587/transinf.2022DLP0058
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E106-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2023
AB - A generative model is presented for two types of person image generation in this paper. First, this model is applied to pose-guided person image generation, i.e., converting the pose of a source person image to the target pose while preserving the texture of that source person image. Second, this model is also used for clothing-guided person image generation, i.e., changing the clothing texture of a source person image to the desired clothing texture. The core idea of the proposed model is to establish the multi-scale correspondence, which can effectively address the misalignment introduced by transferring pose, thereby preserving richer information on appearance. Specifically, the proposed model consists of two stages: 1) It first generates the target semantic map imposed on the target pose to provide more accurate guidance during the generation process. 2) After obtaining the multi-scale feature map by the encoder, the multi-scale correspondence is established, which is useful for a fine-grained generation. Experimental results show the proposed method is superior to state-of-the-art methods in pose-guided person image generation and show its effectiveness in clothing-guided person image generation.
ER -
English
