WDA-Net: Weakly-Supervised Domain Adaptive Segmentation of Electron Microscopy

TL;DR

WDA-Net introduces a weakly-supervised domain adaptation method for electron microscopy segmentation, leveraging sparse annotations and multi-level knowledge to achieve high performance with minimal labeling effort.

Contribution

The paper proposes a novel weakly-supervised domain adaptation framework using sparse point annotations and a task pyramid to improve segmentation across domains.

Findings

Achieves comparable performance to fully supervised models with only 15% point annotations.

Demonstrates robustness to annotation selection and domain shift.

Introduces a cross-position cut-and-paste augmentation to enhance training.

Abstract

Accurate segmentation of organelle instances, e.g., mitochondria, is essential for electron microscopy analysis. Despite the outstanding performance of fully supervised methods, they highly rely on sufficient per-pixel annotated data and are sensitive to domain shift. Aiming to develop a highly annotation-efficient approach with competitive performance, we focus on weakly-supervised domain adaptation (WDA) with a type of extremely sparse and weak annotation demanding minimal annotation efforts, i.e., sparse point annotations on only a small subset of object instances. To reduce performance degradation arising from domain shift, we explore multi-level transferable knowledge through conducting three complementary tasks, i.e., counting, detection, and segmentation, constituting a task pyramid with different levels of domain invariance. The intuition behind this is that after investigating a related source domain, it is much easier to spot similar objects in the target domain than to delineate their fine boundaries. Specifically, we enforce counting estimation as a global constraint to the detection with sparse supervision, which further guides the segmentation. A cross-position cut-and-paste augmentation is introduced to further compensate for the annotation sparsity. Extensive validations show that our model with only 15% point annotations can achieve comparable performance as supervised models and shows robustness to annotation selection.