Major Vessel Segmentation on X-ray Coronary Angiography using Deep Networks with a Novel Penalty Loss Function

Su Yang, Jihoon Kweon, Young-Hak Kim

Keywords: Deep learning, X-ray coronary angiography, Major vessel segmentation, Penalty Loss Function

TL;DR: Fully convolutional networks with a novel loss function showed a good performance in major vessel segmentation of X-ray coronary angiography.

Abstract: In this study, we proposed a segmentation method of major vessels on X-ray coronary angiography using fully convolutional networks based on U-Net architecture. A novel loss function pGD was introduced by adding a term for penalizing false negative and false positive to generalized dice coefficient (GD). DenseNet121 with pGD achieved the highest average DSC of 91.9±8.7%, precision of 91.3±8.8%, and recall of 92.6±9.6%, respectively and showed improved segmentation performance compared to GD.

Code Of Conduct: I have read and accept the code of conduct.

Remove If Rejected: Remove submission from public view if paper is rejected.