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/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Automatic surgical phase recognition is one of the key technologies tosupport Video-Based Assessment (VBA) systems for surgical education. Utilizingtemporal information is crucial for surgical phase recognition, hence variousrecent approaches extract frame-level features to conduct full video temporalmodeling. For better temporal modeling, we propose SlowFast Temporal ModelingNetwork (SF-TMN) for surgical phase recognition that can not only achieveframe-level full video temporal modeling but also achieve segment-level fullvideo temporal modeling. We employ a feature extraction network, pre-trained onthe target dataset, to extract features from video frames as the training datafor SF-TMN. The Slow Path in SF-TMN utilizes all frame features for frametemporal modeling. The Fast Path in SF-TMN utilizes segment-level featuressummarized from frame features for segment temporal modeling. The proposedparadigm is flexible regarding the choice of temporal modeling networks. Weexplore MS-TCN and ASFormer models as temporal modeling networks and experimentwith multiple combination strategies for Slow and Fast Paths. We evaluateSF-TMN on Cholec80 surgical phase recognition task and demonstrate that SF-TMNcan achieve state-of-the-art results on all considered metrics. SF-TMN withASFormer backbone outperforms the state-of-the-art Not End-to-End(TCN) methodby 2.6% in accuracy and 7.4% in the Jaccard score. We also evaluate SF-TMN onaction segmentation datasets including 50salads, GTEA, and Breakfast, andachieve state-of-the-art results. The improvement in the results shows thatcombining temporal information from both frame level and segment level byrefining outputs with temporal refinement stages is beneficial for the temporalmodeling of surgical phases.

SF-TMN: SlowFast Temporal Modeling Network for Surgical Phase Recognition