The Research Of The Recognition And Classification Of The Leukocyte Images

The classification of leucocyte is one item of the blood test. Using the quantitative analysis to analyze the situation of patients'leucocyte, doctors can know the pathogenesis of illnesses patients got and take effective measures in diagnosis and treatment for patients. The quality and efficiency of leucocyte test can be improved by using the digital image processing and pattern recognition. THSI research focuses on the method of automatic recognition and classification based on the characteristics of leukocyte image. THSI method combines image processing, pattern recognition and artificial intelligence. In tHSI method, images captured by microscopes need to be changed into digital images, then they can be automatically recognized by using image segmentation and feature extraction based on the characteristics of leukocyte image.THSI paper mainly introduces the theory of classification and recoginition of leucocyte image which is based on the fuzzy pattern recognition technology. An image recognition system based on the vision system of microscopic image and theory of cellular morphology has been built. In order to prove the feasibility of tHSI research, experiments on the system have been finished.The key point of tHSI research is automatic recognition and classification of leucocyte image. The paper refers to the theories of cell morphology, cell image segmentation and fuzzy pattern recognition.THSI thesis focuses on the following aspects:1. The cellular morphology has been analyzed, and the attributes have been defined based on the rule of cell classification. According to the simulate classification, the model of automatic classification of leukocyte images has been set up.2. The digital processing has been adopted to process the leucocyte images. The automatic segmentation has been studied in the color space of HSI.3. Membership function and classifier based on the fuzzy pattern recognition have been developed.4. The research and experiments of tHSI system have been finished. The feasibility of the method based on a large number of simulation analysis data has been verified. The system can achieve the requirements in accuracy and speed.