English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the zaia usage as premium feature

ZAIA

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the access of premium content as premium feature

Premium Content

Zendy Tools

Zendy Open

The segmentation of coronary arteries in X-ray angiograms by convolutionalneural networks (CNNs) is promising yet limited by the requirement of preciselyannotating all pixels in a large number of training images, which is extremelylabor-intensive especially for complex coronary trees. To alleviate the burdenon the annotator, we propose a novel weakly supervised training framework thatlearns from noisy pseudo labels generated from automatic vessel enhancement,rather than accurate labels obtained by fully manual annotation. A typicalself-paced learning scheme is used to make the training process robust againstlabel noise while challenged by the systematic biases in pseudo labels, thusleading to the decreased performance of CNNs at test time. To solve thisproblem, we propose an annotation-refining self-paced learning framework(AR-SPL) to correct the potential errors using suggestive annotation. Anelaborate model-vesselness uncertainty estimation is also proposed to enablethe minimal annotation cost for suggestive annotation, based on not only theCNNs in training but also the geometric features of coronary arteries deriveddirectly from raw data. Experiments show that our proposed framework achieves1) comparable accuracy to fully supervised learning, which also significantlyoutperforms other weakly supervised learning frameworks; 2) largely reducedannotation cost, i.e., 75.18% of annotation time is saved, and only 3.46% ofimage regions are required to be annotated; and 3) an efficient interventionprocess, leading to superior performance with even fewer manual interactions.

Weakly Supervised Vessel Segmentation in X-ray Angiograms by Self-Paced Learning from Noisy Labels with Suggestive Annotation