Abstract
The picture of the stomach obtained by TV-endoscope system (Toshiba-Machida) was processed by means of histogram flattening of saturation in HSV color space, for the purpose of the improvement of detectability of small, flat lesions with minute changes in color. With the present method of image processing, subtle changes in color were distinctly enhanced, namely, the pale area becoming more pale and the red more red. As a fundamental study, the RGB elements were converted into HSV color space, being separated into hue, value and saturation elements, respectively. Then, the histogram of each separated element was analized. Consequently, only the histogram of saturation showed two peaks. From these evidences, it should be assumed that the changes in color of the gastric mucosa are ascribed to the difference of saturation. Pictures of a lesion of adenoma and two of early gastric cancer were processed through the histogram flattening in HSV color space. Each processed image of the lesion showed the strongly intensified color contrast to the surrounding normal mucosa. From these studies, it is concluded that an image processing of the histogram flatten-ing of saturation in HSV color spece will make it easier to detect flat and small gastric lesions with minute color changes.
