20 March 2015 Investigation of optimal feature value set in false positive reduction process for automated abdominal lymph node detection method
Author Affiliations +
Abstract
This paper presents an investigation of optimal feature value set in false positive reduction process for the automated method of enlarged abdominal lymph node detection. We have developed the automated abdominal lymph node detection method to aid for surgical planning. Because it is important to understand the location and the structure of an enlarged lymph node in order to make a suitable surgical plan. However, our previous method was not able to obtain the suitable feature value set. This method was able to detect 71.6% of the lymph nodes with 12.5 FPs per case. In this paper, we investigate the optimal feature value set in the false positive reduction process to improve the method for automated abdominal lymph node detection. By applying our improved method by using the optimal feature value set to 28 cases of abdominal 3D CT images, we detected about 74.7% of the abdominal lymph nodes with 11.8 FPs/case.
© (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yoshihiko Nakamura, Yoshihiko Nakamura, Yukitaka Nimura, Yukitaka Nimura, Takayuki Kitasaka, Takayuki Kitasaka, Shinji Mizuno, Shinji Mizuno, Kazuhiro Furukawa, Kazuhiro Furukawa, Hidemi Goto, Hidemi Goto, Michitaka Fujiwara, Michitaka Fujiwara, Kazunari Misawa, Kazunari Misawa, Masaaki Ito, Masaaki Ito, Shigeru Nawano, Shigeru Nawano, Kensaku Mori, Kensaku Mori, } "Investigation of optimal feature value set in false positive reduction process for automated abdominal lymph node detection method", Proc. SPIE 9414, Medical Imaging 2015: Computer-Aided Diagnosis, 94143N (20 March 2015); doi: 10.1117/12.2082500; https://doi.org/10.1117/12.2082500
PROCEEDINGS
7 PAGES


SHARE
Back to Top