Deep image-to-image recurrent network with shape basis for automatic vertebra labeling in large-scale 3D CT volumes

US10366491B2 · kind B2 · utility

Assignee

• Siemens Healthcare Limited · DE

Inventors

• Dong Yang · Somerville, US
• Tao Xiong · Shanghai, CN
• Daguang Xu · Princeton, US
• Shaohua Kevin Zhou · Plainsboro, US
• Mingqing Chen · Plainsboro, US
• Zhoubing Xu · Plainsboro, US
• Dorin Comaniciu · Princeton Junction, US
• Jin-hyeong Park · Princeton, US

Key dates

• Filing date: Feb 2, 2018
• Grant date: Jul 30, 2019
• Expiry date: Apr 7, 2038

Classification

• Technology area (CPC G): Physics
• CPC primary: G06V2201/03
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

A method and apparatus for automated vertebra localization and identification in a 3D computed tomography (CT) volumes is disclosed. Initial vertebra locations in a 3D CT volume of a patient are predicted for a plurality of vertebrae corresponding to a plurality of vertebra labels using a trained deep image-to-image network (DI2IN). The initial vertebra locations for the plurality of vertebrae predicted using the DI2IN are refined using a trained recurrent neural network, resulting in an updated set of vertebra locations for the plurality of vertebrae corresponding to the plurality of vertebrae labels. Final vertebra locations in the 3D CT volume for the plurality of vertebrae corresponding to the plurality of vertebra labels are determined by refining the updated set of vertebra locations using a trained shape-basis deep neural network.