Multi-level Asymmetric Contrastive Learning for Volumetric Medical Image Segmentation Pre-training

TL;DR

This paper introduces MACL, a multi-level asymmetric contrastive learning framework that pre-trains encoder and decoder simultaneously, leveraging multi-scale representations to improve volumetric medical image segmentation accuracy.

Contribution

The paper proposes a novel multi-level asymmetric contrastive learning approach that pre-trains encoder and decoder together, utilizing multi-scale features for better segmentation performance.

Findings

Outperforms 11 existing contrastive learning strategies on 8 datasets.

Achieves up to 7.87% higher Dice scores on challenging datasets.

Demonstrates strong generalization across different U-Net variants.

Abstract

Medical image segmentation is a fundamental yet challenging task due to the arduous process of acquiring large volumes of high-quality labeled data from experts. Contrastive learning offers a promising but still problematic solution to this dilemma. Firstly existing medical contrastive learning strategies focus on extracting image-level representation, which ignores abundant multi-level representations. Furthermore they underutilize the decoder either by random initialization or separate pre-training from the encoder, thereby neglecting the potential collaboration between the encoder and decoder. To address these issues, we propose a novel multi-level asymmetric contrastive learning framework named MACL for volumetric medical image segmentation pre-training. Specifically, we design an asymmetric contrastive learning structure to pre-train encoder and decoder simultaneously to provide better initialization for segmentation models. Moreover, we develop a multi-level contrastive learning strategy that integrates correspondences across feature-level, image-level, and pixel-level representations to ensure the encoder and decoder capture comprehensive details from representations of varying scales and granularities during the pre-training phase. Finally, experiments on 8 medical image datasets indicate our MACL framework outperforms existing 11 contrastive learning strategies. i.e. Our MACL achieves a superior performance with more precise predictions from visualization figures and 1.72%, 7.87%, 2.49% and 1.48% Dice higher than previous best results on ACDC, MMWHS, HVSMR and CHAOS with 10% labeled data, respectively. And our MACL also has a strong generalization ability among 5 variant U-Net backbones. Our code will be released at https://github.com/stevezs315/MACL.