Media.Player

Lung Nodule Detection Method in Computed Tomography Images

详细技术说明

None

*Abstract

The inventors have developed a new computerized scheme to automatically detect lung nodules in computed tomography (CT) images. The detection procedure is conducted by analyzing the shape features of the signed distance field of the CT images. A free-response receiver operating characteristics (FROC) curve is computed to assess scheme performance. Performance test on 52 low-dose CT screening examinations that depict 184 verified lung nodules showed that this scheme achieved an asymptotic maximum sensitivity of 95.1% (175/184). The nine missed nodules include two small solid nodules and seven non-solid nodules. The scheme achieved a sensitivity level of 81.7% with 6.5 false-positives per CT scan. This preliminary study demonstrated the feasibility of applying a simple and robust geometric model using the signed distance field to identify suspicious lung nodules. This method can be easily implemented as the first step of any computer-aided detection schemes and has the potential for improving overall detection performance and scheme efficiency. Applications1) Diagnosis of lung cancerAdvantages1) Easy to implementStage of DevelopmentMethod has been tested on lung CT images

*Principal Investigation

Name: Jiantao Pu, Research Assistant Professor

Department: Med-Radiology

国家/地区

美国