The Application Of Fuzzy Logic Theory In Edge Detection Of Erythrocyte

Erythrocyte morphology is an important part in the area of cytology,pathology and hemorheology. Erythrocyte has 95% part of total blood cells and the ability of deformation of erythrocyte will produce parenchyma influence on the characteristic of hemorheology during blood circulation. In some pathology entironment(e.g. hemolytic anemia, pre-leukemia), erythrocyte and its film-structure will change because of various factors. And it is very important to describe erythrocyte conformation and classification in hemorheology characteristic and causes of some diseases.One of the main study area in erythrocyte morphology is the edge detection of erythrocyte and it is the base to the study of erythrocyte morphology. At present, the methods of edge detection of erythrocyte are using traditional theory such as edge one-order or second-order orient differential coefficient and using simple methods to detect edge.But because of the fuzziness of the edge of erythrocyte, the classical methods are mainly focused on the gray-level variety in circumference area of every point. And according to variety of the first order or second order direction differential coefficient, we can detect the edge by some simple methods.Due to the fuzziness of the image itself, the classical methods are time-wasted and the edge is not clear. In order to deal with these problems, the author puts forward a new edge-detection methods based on the fuzzy logic theory. This approach can be summarized as follows: firstly, a threshold is selected in an image by using an automatic algorithm of thresholding; secondly, in order to extract the fuzzy property plane from the spatial domain, a new membership function based on the threshold is defined; thirdly, the modified spatial domain is obtained; finally, the edge detection is executed using min or max operators.All the algorithm have been programmed in the programming platform of Visual C++6.0 in Windows operation system, and they were tested in several images. The results show that the algorithm studied in this paper have high performance than those used before.