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

This paper investigates the automatic monitoring of tool usage during asurgery, with potential applications in report generation, surgical trainingand real-time decision support. Two surgeries are considered: cataract surgery,the most common surgical procedure, and cholecystectomy, one of the most commondigestive surgeries. Tool usage is monitored in videos recorded either througha microscope (cataract surgery) or an endoscope (cholecystectomy). Followingstate-of-the-art video analysis solutions, each frame of the video is analyzedby convolutional neural networks (CNNs) whose outputs are fed to recurrentneural networks (RNNs) in order to take temporal relationships between eventsinto account. Novelty lies in the way those CNNs and RNNs are trained.Computational complexity prevents the end-to-end training of "CNN+RNN" systems.Therefore, CNNs are usually trained first, independently from the RNNs. Thisapproach is clearly suboptimal for surgical tool analysis: many tools are verysimilar to one another, but they can generally be differentiated based on pastevents. CNNs should be trained to extract the most useful visual features incombination with the temporal context. A novel boosting strategy is proposed toachieve this goal: the CNN and RNN parts of the system are simultaneouslyenriched by progressively adding weak classifiers (either CNNs or RNNs) trainedto improve the overall classification accuracy. Experiments were performed in adataset of 50 cataract surgery videos and a dataset of 80 cholecystectomyvideos. Very good classification performance are achieved in both datasets:tool usage could be labeled with an average area under the ROC curve of $A_z =0.9961$ and $A_z = 0.9939$, respectively, in offline mode (using past, presentand future information), and $A_z = 0.9957$ and $A_z = 0.9936$, respectively,in online mode (using past and present information only).

Monitoring tool usage in surgery videos using boosted convolutional and recurrent neural networks