Seung Yeon Shin

TL;DR: In this paper, a graph neural network was incorporated into a unified CNN architecture to exploit both local appearances and global vessel structures for vessel segmentation, and the proposed method outperformed or is on par with current state-of-theart methods in terms of the average precision and the area under the receiver operating characteristic curve.

TL;DR: Experimental results show that the proposed method can successfully localize and classify masses with less annotation effort, and is comparable to results trained from 800 strongly annotated images.

TL;DR: In this article, a weakly and semi-supervised training scenario with appropriate training loss selection was proposed to localize and classify masses in breast ultrasound images, which achieved a 4.5% point increase in CorLoc.

TL;DR: Quantitative and qualitative evaluation conducted on a dataset of 18 sequences demonstrate the effectiveness of the proposed method to extract coronary vessels from fluoroscopic x-ray sequences.

TL;DR: In this article, a cylindrical topological constraint based on persistent homology is applied for small bowel segmentation, where the inner cylinder of the small bowel is free of the touching issue, and the shape constraint on this augmented branch guides the network to generate a topologically correct segmentation.