Automatic image segmentation is vital for the computer-aided determination of treatment directions, particularly in terms of labelling lesions or infected areas. However, the manual labelling of disease regions is inconsistent and a time-consuming assignment. Meanwhile, radiologists’ comments are exceedingly subjective, regularly impacted by personal clinical encounters. To address these issues, we proposed a transformer learning strategy to automatically recognize infected areas in medical images. We firstly utilize a parallel partial decoder to aggregate high-level features and then generate a global feature map. Explicit edge attention and implicit reverse attention are applied to demonstrate boundaries and enhance their expression. Additionally, to alleviate the need for extensive labeled data, we propose a segmentation network combining propagation and transformer architectures that requires only a small amount of labeled data while leveraging fundamentally unlabeled images. The attention mechanisms are integrated within convolutional networks, keeping their global structures intact. Standalone transformers connected straightforwardly and receiving image patches can also achieve impressive segmentation performance. Our network enhanced the learning ability and attained a higher quality execution. We conducted a variety of ablation studies to demonstrate the adequacy of each modelling component. Experiments conducted across various medical imaging modalities illustrate that our model beats the most popular segmentation models. The comprehensive results also show that our transformer architecture surpasses established frameworks in accuracy while better preserving the natural variations in anatomy. Both quantitatively and qualitatively, our model achieves a higher overlap with ground truth segmentations and improved boundary adhesion.

As a crucial task for disease diagnosis, existing semi-supervised segmentation approaches process labeled and unlabeled data separately, ignoring the relationships between them, thereby limiting further performance improvements. In this work, we introduce a transformer-based multi-task framework that concurrently leverages both labeled and unlabeled volumes by encoding shared representation patterns. We first integrate transformers into YOLOv5 to enhance segmentation capabilities and adopt a multi-task approach spanning shadow region detection and boundary localization. Subsequently, we leverage the mean teacher model to simultaneously learn from labeled and unlabeled inputs alongside orthogonal view representations, enabling our approach to harness all available annotations. Our network can improve the learning ability and attain superior performance. Extensive experiments demonstrate that the transformer-powered architecture encodes robust inter-sample relationships, unlocking substantial performance gains by capturing shared information between labeled and unlabeled data. By treating both data types concurrently and encoding their shared patterns, our framework addresses the limitations of existing semi-supervised approaches, leading to improved segmentation accuracy and robustness.