Voxel-level Siamese Representation Learning for Abdominal Multi-Organ Segmentation

•We propose simple yet effective voxel-level representation learning method for multi-organ segmentation on abdominal CT scans. Our method enforces voxel-level feature relations in the representation space so that we can enhance representation power of the base network•We define voxel-level feature relations without using negative samples, which is an efficient method in terms of the computational cost. While using SimSiam method, we neither use a large batch size nor use a momentum encoder, which are typically required for collecting a large amount of negative samples.•We propose a multi-resolution context aggregation method that aggregates features from the intermediate layers and the last hidden layer. Using our method, we can train both global and local context features simultaneously. Background and Objective: Recent works in medical image segmentation have actively explored various deep learning architectures or objective functions to encode high-level features from volumetric data owing to limited image annotations. However, most existing approaches tend to ignore cross-volume global context and define context relations in the decision space. In this work, we propose a novel voxel-level Siamese representation learning method for abdominal multi-organ segmentation to improve representation space. Methods: The proposed method enforces voxel-wise feature relations in the representation space for leveraging limited datasets more comprehensively to achieve better performance. Inspired by recent progress in contrastive learning, we suppressed voxel-wise relations from the same class to be projected to the same point without using negative samples. Moreover, we introduce a multi-resolution context aggregation method that aggregates features from multiple hidden layers, which encodes both the global and local contexts for segmentation. Results: Our experiments on the multi-organ dataset outperformed the existing approaches by 2% in Dice score coefficient. The qualitative visualizations of the representation spaces demonstrate that the improvements were gained primarily by a disentangled feature space. Conclusion: Our new representation learning method successfully encoded high-level features in the representation space by using a limited dataset, which showed superior accuracy in the medical image segmentation task compared to other contrastive loss-based methods. Moreover, our method can be easily applied to other networks without using additional paramete