Self-Attention Guided CycleGAN for Enhanced Reconstruction of Fibrosis Features in Pathological Images

Cheng Xianda, Nicholas Syn, Cheng Han Ng, Aileen Wee, Gwyneth Soon, HirokazuTakahashi, Ho Khek Yu, Mark Muthiah

Keywords: MASLD, Computational Pathology, CycleGAN, Self-Attention, Virtual Staining

TL;DR: In MASLD, fibrosis severity is crucial. This study proposes ROI-SAGAN, an improved CycleGAN, to convert H&E to MT staining, enhancing fibrosis detail for automated diagnosis.

Confirmation Of Submission Requirements: I submit an abstract. It uses the template provided on the submission page and is no longer than 2 pages.

Submission Number: 120