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 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

Zendy Tools

Zendy Open

Deep learning networks have shown promising performance for accurate objectlocalization in medial images, but require large amount of annotated data forsupervised training, which is expensive and expertise burdensome. To addressthis problem, we present a one-shot framework for organ and landmarklocalization in volumetric medical images, which does not need any annotationduring the training stage and could be employed to locate any landmarks ororgans in test images given a support (reference) image during the inferencestage. Our main idea comes from that tissues and organs from different humanbodies have a similar relative position and context. Therefore, we couldpredict the relative positions of their non-local patches, thus locate thetarget organ. Our framework is composed of three parts: (1) A projectionnetwork trained to predict the 3D offset between any two patches from the samevolume, where human annotations are not required. In the inference stage, ittakes one given landmark in a reference image as a support patch and predictsthe offset from a random patch to the corresponding landmark in the test(query) volume. (2) A coarse-to-fine framework contains two projectionnetworks, providing more accurate localization of the target. (3) Based on thecoarse-to-fine model, we transfer the organ boundingbox (B-box) detection tolocating six extreme points along x, y and z directions in the query volume.Experiments on multi-organ localization from head-and-neck (HaN) CT volumesshowed that our method acquired competitive performance in real time, which ismore accurate and 10^5 times faster than template matching methods with thesame setting. Code is available: https://github.com/LWHYC/RPR-Loc.

Contrastive Learning of Relative Position Regression for One-Shot Object Localization in 3D Medical Images