Learning better contrastive view from radiologist's gaze

Open Webpage

Sheng Wang, Zihao Zhao, Zixu Zhuang, Xi Ouyang, Lichi Zhang, Zheren Li, Chong Ma, Tianming Liu, Dinggang Shen, Qian Wang

Published: 01 Jan 2025, Last Modified: 14 May 2025Pattern Recognit. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•This paper extends our conference paper presented at the NeurIPS 2023 Workshop on Gaze Meets ML.•We conduct a thorough analysis of limitations in current data augmentation methods for medical images.•For the first time, we suggest leveraging the gaze of radiologists to craft semantic-aware augmentation during contrastive learning.•We compare human experts’ visual attention and network’s attention. And evaluate them in downstream tasks.