Spatial Content Alignment For Pose Transfer

TL;DR

This paper introduces SCAGAN, a novel pose transfer framework that improves content alignment and detail preservation in generated person images by transferring edge content and using a new synthesis module.

Contribution

The paper proposes a new spatial content alignment method and a Content-Style DeBlk to enhance pose transfer quality, addressing geometric mismatch issues.

Findings

Outperforms state-of-the-art methods in quantitative metrics

Produces more realistic and detailed person images

Ablation studies confirm effectiveness of proposed components

Abstract

Due to unreliable geometric matching and content misalignment, most conventional pose transfer algorithms fail to generate fine-trained person images. In this paper, we propose a novel framework Spatial Content Alignment GAN (SCAGAN) which aims to enhance the content consistency of garment textures and the details of human characteristics. We first alleviate the spatial misalignment by transferring the edge content to the target pose in advance. Secondly, we introduce a new Content-Style DeBlk which can progressively synthesize photo-realistic person images based on the appearance features of the source image, the target pose heatmap and the prior transferred content in edge domain. We compare the proposed framework with several state-of-the-art methods to show its superiority in quantitative and qualitative analysis. Moreover, detailed ablation study results demonstrate the efficacy of our contributions. Codes are publicly available at github.com/rocketappslab/SCA-GAN.