Detection And Implementation Of Microvessel In Living Body Based On Improved Fuzzy Cluster Means Algorithm

Scientific research shows that human microcirculation can be expressed by the diameter of microvessels in conjunctiva and the velocity of bl ood flow in microvessels.Many medical studies have also found that some cardiovascular and cerebrovascular diseases are closely related to the diameter of microvessels in patients and the velocity of blood flow in microvessels.Therefore,the observation and study of microvascular status provide a diagnostic method for some diseases.At the same time,the mechanism analysis of drug action provides direct observation data,which promotes the development of new drugs.In this paper,a microvascular detection system in vivo is studied.It is hoped that a clearer segmentation image can be obtained through this system,and some basic contributions can be made to medical microvascular image analysis.According to the requirement of in vivo observation,this paper designs the magnification,resolution and field of view diameter of microvascular image acquisition system,chooses stereo microscope and CCD industrial digital camera as image acquisition system,and chooses high-brightness LED cold light source and semi-scattering falling light matching system as the light source and matching system of the system;secondly,collect and record the video of microvessel;find the clearest part of microvessel video,cut out 20 sequence images,and intercept the sequence images by window cutting.Then,the microvascular sequence image is stretched separately under RGB color channel,and the weight coefficient of RGB is changed into gray image.The gray image is stretched further,and the gray image is registered.The 20 serial images after registration are segmented by improved FCM algorithm,and 20 frames are fused.At last,the clustering image is binarized to get the segmentation image of microvessel,the interface of microvessel detection system is designed,the experimental process and phenomena are given,and the experimental results are analyzed to prove the feasibility of the system.In order to prove the advancement of the improved FCM algorithm,this paper uses the improved FCM fuzzy clustering algorithm and K-Means clustering algorithm,SOM neural network algorithm and FCM fuzzy clustering algorithm to segment the same microvascular image,and compares the contour effect of the microvascular image after segmentation by four methods.Experiments show that the improved FCM algorithm is more accurate than K-Means algorithm,SOM algorithm and FCM fuzzy clustering algorithm in the segmentation of microvessels in vivo,and the connectivity of microvessels is stronger.