AdaptFRCNet: Semi-supervised adaptation of pre-trained model with frequency and region consistency for medical image segmentation

Along He, Yanlin Wu, Zhihong Wang, Tao Li, Huazhu Fu

Abstract: Highlights•For PEFT, we propose lightweight attention adapter with a few trainable parameters to adapt large pre-trained models to the medical field, and it significantly reduces computational costs while leveraging the powerful feature representation capabilities of large pre-trained models.•We introduce frequency domain consistency (FDC) and multi-granularity region similarity consistency (MRSC) regularization strategies for semi-supervised learning, enabling learning from unlabeled data to address insufficient annotation. MRSC is advantageous for medical image segmentation with diverse shapes and scales, and FDC effectively utilizes frequency signals to enhance segmentation performance.•Our method is plug-and-play and can be seamlessly integrated with existing methods. During inference, FDC and MRSC can be safely removed, thus further reducing inference complexity. Finally, our method outperforms the current state-of-the-art (SOTA) semi-supervised methods on three publicly available medical datasets.