Weakly-Supervised Cross-Domain Adaptation for Endoscopic Lesions Segmentation

TL;DR

This paper introduces a weakly-supervised domain adaptation framework for endoscopic lesion segmentation that leverages transferable knowledge and self-supervised pseudo labels to improve accuracy while reducing annotation costs.

Contribution

It proposes a novel transferability framework and pseudo label generator to effectively transfer knowledge across datasets and prevent negative transfer in lesion segmentation.

Findings

Outperforms existing methods on endoscopic and public datasets.

Effectively explores transferable domain-invariant features.

Reduces false pseudo labels through self-supervision.

Abstract

Weakly-supervised learning has attracted growing research attention on medical lesions segmentation due to significant saving in pixel-level annotation cost. However, 1) most existing methods require effective prior and constraints to explore the intrinsic lesions characterization, which only generates incorrect and rough prediction; 2) they neglect the underlying semantic dependencies among weakly-labeled target enteroscopy diseases and fully-annotated source gastroscope lesions, while forcefully utilizing untransferable dependencies leads to the negative performance. To tackle above issues, we propose a new weakly-supervised lesions transfer framework, which can not only explore transferable domain-invariant knowledge across different datasets, but also prevent the negative transfer of untransferable representations. Specifically, a Wasserstein quantified transferability framework is developed to highlight widerange transferable contextual dependencies, while neglecting the irrelevant semantic characterizations. Moreover, a novel selfsupervised pseudo label generator is designed to equally provide confident pseudo pixel labels for both hard-to-transfer and easyto-transfer target samples. It inhibits the enormous deviation of false pseudo pixel labels under the self-supervision manner. Afterwards, dynamically-searched feature centroids are aligned to narrow category-wise distribution shift. Comprehensive theoretical analysis and experiments show the superiority of our model on the endoscopic dataset and several public datasets.