IDEAL: Improved DEnse locAL Contrastive Learning for Semi-Supervised Medical Image Segmentation

TL;DR

IDEAL introduces a dense contrastive learning approach with a novel loss and regularization mechanism, significantly improving semi-supervised medical image segmentation performance, especially in cardiac MRI tasks.

Contribution

The paper presents a new dense pixel-level contrastive loss and a bidirectional consistency regularization for semi-supervised medical image segmentation.

Findings

Outperforms state-of-the-art methods on cardiac MRI segmentation.

Effective in leveraging limited labeled data.

Improves local feature representations in dense segmentation tasks.

Abstract

Due to the scarcity of labeled data, Contrastive Self-Supervised Learning (SSL) frameworks have lately shown great potential in several medical image analysis tasks. However, the existing contrastive mechanisms are sub-optimal for dense pixel-level segmentation tasks due to their inability to mine local features. To this end, we extend the concept of metric learning to the segmentation task, using a dense (dis)similarity learning for pre-training a deep encoder network, and employing a semi-supervised paradigm to fine-tune for the downstream task. Specifically, we propose a simple convolutional projection head for obtaining dense pixel-level features, and a new contrastive loss to utilize these dense projections thereby improving the local representations. A bidirectional consistency regularization mechanism involving two-stream model training is devised for the downstream task. Upon comparison, our IDEAL method outperforms the SoTA methods by fair margins on cardiac MRI segmentation. Code available: https://github.com/hritam-98/IDEAL-ICASSP23