SEAN: image synthesis with semantic region-adaptive normalization

Zhu, Peihao, Abdal, Rameen, Qin, Yipeng ORCID: https://orcid.org/0000-0002-1551-9126 and Wonka, Peter 2020. SEAN: image synthesis with semantic region-adaptive normalization. Presented at: Conference on Computer Vision and Pattern Recognition (CVPR 2020), Seattle, Washington, USA, 14-19 June 2020.

Preview	PDF - Accepted Post-Print Version Download (7MB) | Preview
Preview	PDF - Supplemental Material Download (19MB) | Preview

Abstract

We propose semantic region-adaptive normalization (SEAN),asimplebuteffectivebuildingblockforGenerative Adversarial Networks conditioned on segmentation masks that describe the semantic regions in the desired output image. Using SEAN normalization, we can build a network architecture that can control the style of each semantic region individually, e.g., we can specify one style reference image per region. SEAN is better suited to encode, transfer, and synthesize style than the best previous method in terms of reconstruction quality, variability, and visual quality. We evaluateSEANonmultipledatasetsandreportbetterquan titative metrics (e.g. FID, PSNR) than the current state of the art. SEAN also pushes the frontier of interactive image editing. We can interactively edit images by changing segmentation masks or the style for any given region. We can also interpolate styles from two reference images per region. Code: https://github.com/ZPdesu/SEAN.

Item Type:	Conference or Workshop Item (Paper)
Status:	In Press
Schools:	Computer Science & Informatics
Date of First Compliant Deposit:	5 May 2020
Date of Acceptance:	27 February 2020
Last Modified:	07 Nov 2022 10:11
URI:	https://orca.cardiff.ac.uk/id/eprint/131426

Citation Data

Cited 123 times in Scopus. View in Scopus. Powered By Scopus® Data

Actions (repository staff only)

Edit Item