Fourier Semi-Supervised Learning Method for Medical Image Segmentation
The scarcity of labeled data is a challenging problem that affects the segmentation accuracy of medical images.Aiming to solve this problem,a new semi-supervised learning method based on Fourier transform and consistent constraintis proposed.In the case of a small amount of annotated data,the output of unannotated data via Fourier transform interpolation and model segmentation is spatially consistent with the output of reverse operation,and the consistency regularization constraint for unannotated data is constructed to improve the model performance of fully supervised learning.The experimental results,based on the openly available datasets ofthe automatic cardiac diagnosis challenge,synapse and computed tomography lymph node,demonstratethat the proposed algorithm is superior to baseline methods and can be integrated with existing state-of-the-art semi-supervised medical image segmentation methods to improve their segmentation performances.
medical image segmentationsemi-supervised learningFourier transformationconsistency regularization constraint
NETL
NSTL
万方数据
