Realtime RGB-subtraction image processing.

概要

論文の詳細を見る
We designed a system with an RGB (Red, Green, Blue) colour subtraction unit. RGB-subtraction image processing was used before or/and after scattering methylene blue dye for various gastric lesions. As for a preliminary experiment, we measured the refractory characteristic curve of normal gastric mucosa after scattering the methylene blue and obtained the mean refractory peak wave length. We studied the image gained by using the RGB-subtraction processing unit and investigated the subtraction rate and gain number as parameters. From the preliminary experiment, the mean peak value of the reflectory rate was 452 nm and in the blue colour zone. The R-G image (ie, that produced by subtracting red and green) was similar to the R -B image as the G-R image did the B-R image and the G-B image the B-G image after RGB-subtracting for the original image. RGB-subtraction image processing showed us the border between diseased and normal mucosa and mucosal changes clearly after methylene blue dye scattering. The border between normal and lesioned mucosa on these images corresponded to those prepared histologically. This image processing enables us to detect the lesions more clearly and the lesional area on gastric mucosa accurately. TV-endoscopy using RGB-subtraction image processing may permit endoscopic surgery.
社団法人日本消化器内視鏡学会の論文