L-SNet: from Region Localization to Scale Invariant Medical Image Segmentation

TL;DR

This paper introduces L-SNet, a differentiable two-stage network for medical image segmentation that improves accuracy and consistency in handling large scale variations by integrating localization and segmentation with a RoI recalibration module.

Contribution

The paper proposes a novel differentiable two-stage architecture with a RoI recalibration module, addressing performance bottlenecks and training inconsistencies in existing coarse-to-fine segmentation models.

Findings

Outperforms state-of-the-art coarse-to-fine models on public datasets.

Maintains high accuracy with negligible additional computation.

Addresses training and performance issues in large scale variation segmentation.

Abstract

Coarse-to-fine models and cascade segmentation architectures are widely adopted to solve the problem of large scale variations in medical image segmentation. However, those methods have two primary limitations: the first-stage segmentation becomes a performance bottleneck; the lack of overall differentiability makes the training process of two stages asynchronous and inconsistent. In this paper, we propose a differentiable two-stage network architecture to tackle these problems. In the first stage, a localization network (L-Net) locates Regions of Interest (RoIs) in a detection fashion; in the second stage, a segmentation network (S-Net) performs fine segmentation on the recalibrated RoIs; a RoI recalibration module between L-Net and S-Net eliminating the inconsistencies. Experimental results on the public dataset show that our method outperforms state-of-the-art coarse-to-fine models with negligible computation overheads.