Study On MR Image Analysis

Recent years, as the development of medical image technology, medical image analysis is becoming more and more useful in clinical diagnoses and researches. Modern medicine has become more and more dependent on the information from medical image. At the same time, computer technology and graphics and image technology has been highly developed. That made medical image and computer combined tightly, called digital medicine. For the help of graphics and image technology, the quality and the approach of display has been exceedingly improved. The level of diagnosis and treat has also been improved.MRI ( Magnetic Resonance Imaging ) is one of the most important imaging approaches. Because MR images has high tissue resolution and the anatomical structures can be displayed clearly. MRI has extremely advantages in the diagnosis of Central Nervous System Diseases. This thesis presents a medical image based method of focus extraction and medical parameter calculation of brain MR images.First, the image dataset is obtained from DICOM files, which are standard medical images. Then the focus is separated from normal tissues by the thresholding segmentation method. The threshold is automatic selected by maximum entropy method and minimum cross entropy method, which is optimized by GA (Genetic Algorithm). Then a templet of the skull and other tissues (skin, fat) is obtained from a series of normal brain MR images. Then the focus can be extracted from the images, using the templet. Finally, some important medical parameters are calculated, such as the area and the position of the focus.The results of experiments show the method presented in this thesis is effective to extract focus from MR images and can calculated important medical parameters which are useful for medical diagnosis.