부산대학교 도서관

To improve the efficiency of image analysis of automatic blood type reading system, a new feature extraction in image processing algorithms is presented. Image processing algorithm based on color channel separation and entropy processing is therefore proposed to extract the color characteristics and gray scale characteristics of blood type images. Under the circumstance of the red blood cell reaction phenomenon with complex diversity, the novel feature extraction algorithm adopts comprehensive extraction of color and gray scale features, which has a greater effect on the efficiency of image analysis than individual feature analysis. Compare to the traditional gray scale projection curve algorithm, the feature extraction of this system basically complies with the requirements of blood type reading.