Authors: Dan Song, Xuanpu Zhang, Juan Zhou, Weizhi Nie, Ruofeng Tong, Mohan Kankanhalli, An-An Liu

Published on: November 08, 2023

Impact Score: 7.8

Arxiv code: Arxiv:2311.04811

Summary

• What is new: Introduces a comprehensive analysis of image-based virtual try-on technologies, highlights gaps for commercial use, and proposes semantic criteria with CLIP for evaluation.
• Why this is important: The gap between research progress in image-based virtual try-on and its commercial application, plus a lack of comprehensive overview in this field.
• What the research proposes: A detailed survey of current techniques and methodologies, introduction of a new semantic criteria for evaluation, and provision of uniformly implemented evaluation metrics and dataset.
• Results: Identification of unresolved issues, future research directions, and a public dataset with evaluation metrics for further exploration.

Technical Details

Technological frameworks used: CLIP for semantic criteria

Models used: State-of-the-art techniques in pipeline architecture, person representation, try-on indication, clothing warping, and try-on stage.

Data used: Uniformly implemented evaluation metrics on the same dataset

Potential Impact

Online shopping platforms and fashion retailers could benefit or be disrupted by the advancements in image-based virtual try-on technologies.

Want to implement this idea in a business?

We have generated a startup concept here: Virtual FitMirror.