A proposal of imaging analysis method for accurate extraction of capillary areas using HSV model feature

Purpose of this study is to develop a method for accurate capillary area extraction from a captured image using RGB and HSV features. Experiments were conducted with eleven subjects. The proposed method utilizing OpenCV platform produced clearer capillary images compared to the original images: with capillary information, without capillary information, images with halation, images with black dots. SVM classification of capillary image with acquired numeric features of G, H, S and V values yielded over 90 % accuracy. The results have shown the effectiveness of the proposed method.