Weakly-supervised Medical Image Segmentation with Gaze Annotations

TL;DR

This paper introduces a novel weakly-supervised medical image segmentation method using gaze annotations, leveraging human attention to reduce annotation costs and improve segmentation accuracy, validated on polyp and prostate datasets.

Contribution

It presents a new gaze annotation scheme and a multi-level training framework that effectively utilizes gaze heatmaps and cross-level consistency to enhance segmentation performance.

Findings

Gaze annotations outperform previous label-efficient schemes.

The method achieves high segmentation accuracy with less annotation time.

Introduces the first gaze dataset for medical image segmentation.

Abstract

Eye gaze that reveals human observational patterns has increasingly been incorporated into solutions for vision tasks. Despite recent explorations on leveraging gaze to aid deep networks, few studies exploit gaze as an efficient annotation approach for medical image segmentation which typically entails heavy annotating costs. In this paper, we propose to collect dense weak supervision for medical image segmentation with a gaze annotation scheme. To train with gaze, we propose a multi-level framework that trains multiple networks from discriminative human attention, simulated with a set of pseudo-masks derived by applying hierarchical thresholds on gaze heatmaps. Furthermore, to mitigate gaze noise, a cross-level consistency is exploited to regularize overfitting noisy labels, steering models toward clean patterns learned by peer networks. The proposed method is validated on two public medical datasets of polyp and prostate segmentation tasks. We contribute a high-quality gaze dataset entitled GazeMedSeg as an extension to the popular medical segmentation datasets. To the best of our knowledge, this is the first gaze dataset for medical image segmentation. Our experiments demonstrate that gaze annotation outperforms previous label-efficient annotation schemes in terms of both performance and annotation time. Our collected gaze data and code are available at: https://github.com/med-air/GazeMedSeg.