Features for Discriminating Normal Cases in Automated Mass Screening for Gastric Cancer Using Double Contrast X-ray Images of Stomach

Abstract

In a mass screening for gastric cancer, diagnosticians read many stomach X-ray pictures at a time. To decrease the number of reading the pictures in the mass screenings, the proposed method discriminates normal cases in stomach X-ray images using the propposed features. In normal cases, folds on the stomach wall appear in parallel in the images. Therefore, the proposed method measures characteristics of parallelism for the folds in the images. Experimental results of the discriminations for 103 images have shown that the proposed features are well effective for recognizing normal cases.