Feature Extraction And Classification For 2d Medical Image

With the progress of the visualization technology, modern medical technology depends more on the image processing, the visualization technology is playing an important role in clinical diagnosis and science research. One of the important fields is the image classification based on different image feature extraction strategies, which contributes quite a lot to the clinic diagnosis. Characteristic extracting of medical image and mode classification are focus on medical science field, which are not only important in theoretical research, but also in clinical diagnosis and treatment,The paper focus mainly on the subject of feature definition and extraction of medical image prior to the image classification, which includes:1. Techniques of medical image obtaining and pre-processing covering from DICOM standards, image enhancement and geometric correction, to various methods of image optimizing.2. Several most frequently used image feature extraction approaches and dedicated revisions. The features concerned are such as texture feature, gray histogram feature and features derive form the principle component analysis.3. Edge-detection of medical image based on 2d gradient and second-order derivative. Especially a revised primary contour model is provided.4. Three image classification methods based on precise image matching, gray scale histogram analysis, and neural network applying on the features derive form the principle component.5. An executable system named Medical Image Visual System using to verify the principal algorithms mentioned in the thesis, with which several highly detailed experiments as well as the analysis of the results are further presented.