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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Representation learning of the task-oriented attention while trackinginstrument holds vast potential in image-guided robotic surgery. Incorporatingcognitive ability to automate the camera control enables the surgeon toconcentrate more on dealing with surgical instruments. The objective is toreduce the operation time and facilitate the surgery for both surgeons andpatients. We propose an end-to-end trainable Spatio-Temporal Multi-TaskLearning (ST-MTL) model with a shared encoder and spatio-temporal decoders forthe real-time surgical instrument segmentation and task-oriented saliencydetection. In the MTL model of shared parameters, optimizing multiple lossfunctions into a convergence point is still an open challenge. We tackle theproblem with a novel asynchronous spatio-temporal optimization (ASTO) techniqueby calculating independent gradients for each decoder. We also design acompetitive squeeze and excitation unit by casting a skip connection thatretains weak features, excites strong features, and performs dynamic spatialand channel-wise feature recalibration. To capture better long termspatio-temporal dependencies, we enhance the long-short term memory (LSTM)module by concatenating high-level encoder features of consecutive frames. Wealso introduce Sinkhorn regularized loss to enhance task-oriented saliencydetection by preserving computational efficiency. We generate the task-awaresaliency maps and scanpath of the instruments on the dataset of the MICCAI 2017robotic instrument segmentation challenge. Compared to the state-of-the-artsegmentation and saliency methods, our model outperforms most of the evaluationmetrics and produces an outstanding performance in the challenge.

ST-MTL: Spatio-Temporal Multitask Learning Model to Predict Scanpath While Tracking Instruments in Robotic Surgery