article Open AccessTop 1% cited

AnatomyNet: Deep learning for fast and fully automated whole‐volume segmentation of head and neck anatomy

Medical Physics · 2018 · Vol. 46(2) · pp. 576–589

Wentao Zhu(University of California, Irvine)Yufang Huang(Lenovo (China))Liang Zeng(Shanghai Zhaozhan Metal Materials)Xuming Chen(Shanghai Chest Hospital)Yong Liu(Shanghai Jiao Tong University)Zhen Qian(Tencent (China))Nan Du(Tencent (China))Wei Fan(Tencent (China))Xiaohui Xie✉(University of California, Irvine)

View via DOI Publisher page

Abstract

Deep learning models offer a feasible solution to the problem of delineating OARs from CT images. We demonstrate that our proposed model can improve segmentation accuracy and simplify the autosegmentation pipeline. With this method, it is possible to delineate OARs of a head and neck CT within a fraction of a second.

Dental Radiography and Imaging Medical Imaging and Analysis Anatomy and Medical TechnologySegmentationComputer scienceArtificial intelligenceDeep learningAnatomyGround truthOptic chiasmSørensen–Dice coefficientHead and neckFeature (linguistics)