Research On The Automatic Recognition Of Lymphocyte

The microscopic blood-cell-image processing is an important method in clinical inspection. As environmental pollution, chemical waste and the effects of physical factors, the number of people suffering from blood diseases has been increasing in recent years. So using computer-aided recognition is of great significance for the early diagnosis and classification. The conventional approaches are blood cell examination (i.e. number, morphology, etc) from peripheral blood and bone smears by optical microscope. Generally in blood analysis, different categories of blood cells are distinguished by their size and color. This purely visual observation has the problem of poor reproducibility between the observers and themselves. Therefore, it is necessary to develop an automatic cell classifying and diagnostic system to complete a quantitative analysis and computer-aided detection of blood cell. The recognition of cells is the hotspot in the domain of Image Processing and Pattern Recognition, and has sufficiently wide foreground in the domain of Biomedical Application. Through the computer-aided diagnosis, it has provided more abundant and exacter information for medicine.This paper deals with the research in the application of image process and recognition in the field of medical. Based on the 24-bit true color lymphocyte image, this article conducts an in-depth research and study on how to use the technology of computer image processing and recognition to achieve its automatic analysis, in order to reducing the artificial error of judging and the intensity of labour and boosting the work efficiency. According to the characteristics of the lymphocyte in the blood and marrow, all studies focus on the key techniques of automatic recognition and classification for blood cells such as segmentation of cells, feature extraction and the classification technique. The main contents are listed below:(1) Automatic image segmentation is the key step since the results of segmentation directly influences the subsequent feature extraction and recognition. So an automated segmentation algorithm fusing multi-color space is proposed for segmentation of the color lymphocyte image. Through the RGB color space based on Maximum Variance theory, the nucleus is accurate segmented, and then using the color information of the cytoplasm, the cytoplasm is segmented through the HSI color space. In order to detect the nucleus edge and cytoplasm edge, a new method based on the mathematical morphologic and Canny operator is proposed. At last the process and analyse of single lymphocyte can be realized by the arithmetic of auto-detection. The results show that the method is valid and efficient to segment color images from lymphocyte smears.(2) The characteristics of lymphocyte are the basis for their classification. After the cell segmentation from the complex background, the next step is to find distinguished features for analyzing and classification, that is to conduct feature extraction. Based on the segmentation images, combined with the apriori knowledge of the clinical pathologist, the characteristics including morphological features, chrominance and luminance features, color features and texture features are extracted.(3) For the classification, in order to solve the problems of high characteristic dimension and multi-category classification, and to enhance the correct rate of classification, the application of Back Propagation neural network is studied to achieve the automatic classification of the cells and the methods are presented to construct BP classification and select control parameters. In this thesis, the traditional BP classification is improved and the methods cross validation is used to verify the performance of it. The improved BP classification is compared with the traditional BP classification through the results of the recognition and the results indicate that it can be used in classification of lymphocyte with good effect to be realized.Finally, we make a conclusion and propose the future research directions in this field.