Semi-Supervised Medical Image Segmentation with Cross-View Consistency and Contrastive Learning

Abstract

Medical image segmentation plays a crucial role in many clinical applications. To alleviate the dependency on massive annotations, semi-supervised learning has attracted increasing attention. However, these methods face significant intra-class and inter-class variation and do not fully utilize the critical multi-view information inherent in medical images. Therefore, this study proposes a novel network, CV-Net, which integrates multi-view information for semi-supervised medical image segmentation. Concretely, the network is based on Mean-Teacher architecture which largely narrows the empirical distribution gap between labeled and unlabeled data. Cross-view consistency regularization incorporates a dual-branch attention architecture to integrate consistent semantics while focusing on details, enhancing feature extraction capabilities. Bi-semantic contrastive learning leverages limited labels and explores pseudo-labels to define semantically similar regions, enhancing the representation capacity. Compared with the previous state-of-the-art methods, CV-Net demonstrates a notable advantage in two datasets. Specifically, the proposed network reduces both intra-class and inter-class variability with 5% and 10% labeled data.

Note:
Funding Information: This research was funded by the National Natural Science Foundation of China (62172437) and National HighLevel Hospital Clinical Research Funding (2022-PUMCHC-027).

Conflict of Interests: The authors declare that they have no known competing f inancial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethical Approval: The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Peking Union Medical College Hospital (No. K2733).

Keywords: Medical image segmentationSemi-supervised learning Cross-view consistencyContrastive learning