SAMBV: A fine-tuned SAM with interpolation consistency regularization for semi-supervised bi-ventricle segmentation from cardiac MRI.
Authors
Affiliations (7)
Affiliations (7)
- School of Artificial Intelligence, Hebei University of Technology (HeBUT), Tianjin, 300401, China. Electronic address: [email protected].
- Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, 518055, China. Electronic address: [email protected].
- School of Engineering Science, University of Chinese Academy of Sciences, Beijing, 101408, China. Electronic address: [email protected].
- School of Artificial Intelligence, Hebei University of Technology (HeBUT), Tianjin, 300401, China. Electronic address: [email protected].
- School of Artificial Intelligence, Hebei University of Technology (HeBUT), Tianjin, 300401, China. Electronic address: [email protected].
- School of Artificial Intelligence, Hebei University of Technology (HeBUT), Tianjin, 300401, China. Electronic address: [email protected].
- Key Laboratory of Hormones and Development, Chu Hsien-I Memorial Hospital and Tianjin Institute of Endocrinology, Tianjin Medical University, Tianjin, 300134, China. Electronic address: [email protected].
Abstract
The SAM (segment anything model) is a foundation model for general purpose image segmentation, however, when it comes to a specific medical application, such as segmentation of both ventricles from the 2D cardiac MRI, the results are not satisfactory. The scarcity of labeled medical image data further increases the difficulty to apply the SAM to medical image processing. To address these challenges, we propose the SAMBV by fine-tuning the SAM for semi-supervised segmentation of bi-ventricle from the 2D cardiac MRI. The SAM is tuned in three aspects, (i) the position and feature adapters are introduced so that the SAM can adapt to bi-ventricle segmentation. (ii) a dual-branch encoder is incorporated to collect missing local feature information in SAM so as to improve bi-ventricle segmentation. (iii) the interpolation consistency regularization (ICR) semi-supervised manner is utilized, allowing the SAMBV to achieve competitive performance with only 40% of the labeled data in the ACDC dataset. Experimental results demonstrate that the proposed SAMBV achieves an average Dice score improvement of 17.6% over the original SAM, raising its performance from 74.49% to 92.09%. Furthermore, the SAMBV outperforms other supervised SAM fine-tuning methods, showing its effectiveness in semi-supervised medical image segmentation tasks. Notably, the proposed method is specifically designed for 2D MRI data.