1999 Volume 41 Issue 2 Pages 166-172
The aim of the present study was to analyze the recognition of gingival redness using computer-assisted colorimetric image analysis. Nine dentists, who had career as a clinical periodontist for over 10 years, participated in the evaluation of gingiva. Dentists were shown 4 digital gingival images on the monitor, then defined the regions (ROI) from the point of color according to the category. The images used in this study were saved with Tag Image File Format (TIFF), those had 224 color depth (red 28, green 28, blue 28). The size of each image was 2, 268×1, 764 in pixel-number and 11MB in memory size. The category consisted of healthy, slightly inflamed and severely inflamed. The number of pixels and the average trichromatic gray value of the ROI were measured. Statistical significance among 3 groups was analyzed with Student's t-test. Then 1 image of the inflamed gingiva was separated into 3 trichromatic images. The contrast similarity of the trichromatic images to the original color image was investigated. The most available value in trichromatic color for the recognition of gingival redness was that of green (t-value between healthy and slightly inflamed: R; 3.12 (p=2.3×10-4) G; 4.16 (p=6.5×10-5) B; 3.29 (p=1.4× 0-4), t-value between slightly inflamed and severely inflamed: R; 4.45 (p=2×10-5) G; 9.18 (p=-1× 10-9) B; 9.45 (p=-1×10-9)). Of the 3 separated trichromatic images, the most similar image to the original was the green image.
These findings suggest that the green value is important in recognizing the redness of gingiva by dentists. We are going to utilize a modified green value in the development of an algorithm that recognizes the redness of gingiva automatically. J. Jpn. Soc. Periodontol., 41: 166-172, 1999.