Enjoying Information Dividend: Gaze Track-based Medical Weakly Supervised Segmentation

TL;DR

GradTrack introduces a novel gaze-aware framework that leverages detailed physician gaze data to improve weakly supervised medical image segmentation, significantly enhancing accuracy and narrowing the gap with fully supervised methods.

Contribution

The paper presents GradTrack, a new framework that fully exploits gaze data, including fixation points, durations, and order, for improved weakly supervised segmentation in medical imaging.

Findings

Outperforms existing gaze-based methods with Dice improvements of 3.21% and 2.61%.

Narrowing the performance gap with fully supervised models.

Effective multi-level gaze supervision during decoding process.

Abstract

Weakly supervised semantic segmentation (WSSS) in medical imaging struggles with effectively using sparse annotations. One promising direction for WSSS leverages gaze annotations, captured via eye trackers that record regions of interest during diagnostic procedures. However, existing gaze-based methods, such as GazeMedSeg, do not fully exploit the rich information embedded in gaze data. In this paper, we propose GradTrack, a framework that utilizes physicians' gaze track, including fixation points, durations, and temporal order, to enhance WSSS performance. GradTrack comprises two key components: Gaze Track Map Generation and Track Attention, which collaboratively enable progressive feature refinement through multi-level gaze supervision during the decoding process. Experiments on the Kvasir-SEG and NCI-ISBI datasets demonstrate that GradTrack consistently outperforms existing gaze-based methods, achieving Dice score improvements of 3.21\% and 2.61\%, respectively. Moreover, GradTrack significantly narrows the performance gap with fully supervised models such as nnUNet.