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

Accurate detection of thyroid lesions is a critical aspect of computer-aideddiagnosis. However, most existing detection methods perform only one featureextraction process and then fuse multi-scale features, which can be affected bynoise and blurred features in ultrasound images. In this study, we propose anovel detection network based on a feature feedback mechanism inspired byclinical diagnosis. The mechanism involves first roughly observing the overallpicture and then focusing on the details of interest. It comprises two parts: afeedback feature selection module and a feature feedback pyramid. The feedbackfeature selection module efficiently selects the features extracted in thefirst phase in both space and channel dimensions to generate high semanticprior knowledge, which is similar to coarse observation. The feature feedbackpyramid then uses this high semantic prior knowledge to enhance featureextraction in the second phase and adaptively fuses the two features, similarto fine observation. Additionally, since radiologists often focus on the shapeand size of lesions for diagnosis, we propose an adaptive detection headstrategy to aggregate multi-scale features. Our proposed method achieves an APof 70.3% and AP50 of 99.0% on the thyroid ultrasound dataset and meets thereal-time requirement. The code is available athttps://github.com/HIT-wanglingtao/Thinking-Twice.

Thinking Twice: Clinical-Inspired Thyroid Ultrasound Lesion Detection Based on Feature Feedback