The computerized lung nodule detection by the energy difference method using a standard chest image and soft tissue chest image

摘要

(57) [summary] storage medium method for detecting an abnormality in the lung that has been computerized get soft tissue digital chest image [challenge] standard digital chest image and (10), and the system,. (20), generates first and second difference image from the soft tissue digital chest image [SOLUTION] standards, to identify the abnormal candidate first, in the second difference image, abnormality from a first differential image with a standard digital chest image These analyzes the first and second features of extracting first feature that is predetermined for the candidate, extracts the second feature of a predetermined abnormality candidate from the second difference image and the soft tissue digital chest image, the extracted Each artificial neural network features extracted from abnormal candidates that remain after excluding false positive abnormality candidate identify false positive abnormality candidate corresponding to the features and excluded, is derived from the first and second difference image performs a logical OR operation of the abnormal candidate is applied, derived from the first and second difference image, and excludes false positive abnormality candidate further.