Automatic Segmentation Of Cardiac Images Based On Deep Learning

As the development of medical image processing,automatic Cardiac MRI segmentation has been one of the most important issues.However,the unclear boundaries and inhomogeneous intensity in MRI images usually cause indistinction for pixels from different class around boundaries,limiting accuracy of the segmentation result.In this paper,transformer is introduced into UNet,which is common in medical image segmentation to capture local and global information while direction feature is used enhance to comparison for inter-classes pixels for better performance near the boundaries.The network can be divided into segmentation module and direction feature module:self-attention mechanism is used to replace the convolution module in building blocks for long-range dependence;result of segmentation module is taken as input of direction feature module,which utilize direction vector from pixel to its nearest boundary pixel to improve segmentation accuracy nearby boundaries.Regarding of the limited samples in medical image datasets,this paper used cross pseudo supervision to study the ability of the proposed network to be used in semi-supervision area Several semi-supervision methods are used as comparisonPerformance of the network is evaluated on the 2017 MICCAI Automated Cardiac Diagnosis Challenge(ACDC)dataset and Multi-Centre,Multi-Vendor&Multi-Disease Cardiac Image Segmentation Challenge dataset and got satisfied result.