The recognition of platelets adhered images

An in vitro flow system is usually used to investigate the distribution features of adhered platelets under different flow conditions. The four controlled conditions are: (1) fibrin-coated surface, shear rate 445/s; (2) fibrin-coated surface, shear rate 70/s; (3) fibrinogen-coated surface, shear rate 445/s; (4) fibrinogen-coated surface, shear rate 70/s. The purpose of this study is to apply image processing technique for quantitative analysis of the platelets images. The identification of platelets is realized by sequential procedures, including automatic thresholding, unit-platelet models computation and recognition of platelets' centers by comparing correlation coefficients. Based upon the measures of the number of identified platelets and the mass center of those platelets distribution, the accuracy of the platelets identification is more than 92% in average and is also better than the previous method. A simple GUI(Graphic User Interface) is also added for rendering the physiologists a user friendly man-machine interface.