eduzhai > Applied Sciences > Engineering >

Collaborative Boundary-aware Context Encoding Networks for Error Map Prediction

  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: Medical image segmentation is usually regarded as one of the most importantintermediate steps in clinical situations and medical imaging research. Thus,accurately assessing the segmentation quality of the automatically generatedpredictions is essential for guaranteeing the reliability of the results of thecomputer-assisted diagnosis (CAD). Many researchers apply neural networks totrain segmentation quality regression models to estimate the segmentationquality of a new data cohort without labeled ground truth. Recently, a novelidea is proposed that transforming the segmentation quality assessment (SQA)problem intothe pixel-wise error map prediction task in the form ofsegmentation. However, the simple application of vanilla segmentationstructures in medical image fails to detect some small and thin error regionsof the auto-generated masks with complex anatomical structures. In this paper,we propose collaborative boundaryaware context encoding networks called AEP-Netfor error prediction task. Specifically, we propose a collaborative featuretransformation branch for better feature fusion between images and masks, andprecise localization of error regions. Further, we propose a context encodingmodule to utilize the global predictor from the error map to enhance thefeature representation and regularize the networks. We perform experiments onIBSR v2.0 dataset and ACDC dataset. The AEP-Net achieves an average DSC of0.8358, 0.8164 for error prediction task,and shows a high Pearson correlationcoefficient of 0.9873 between the actual segmentation accuracy and thepredicted accuracy inferred from the predicted error map on IBSR v2.0 dataset,which verifies the efficacy of our AEP-Net.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×