Research On Processing Of Computed Tomography Image Of The Brain

Medical imageology has become one of fields on medical technology that develop most quickly for nearly more than 20 years .It has been applied extensively in diagnosis and treatment,and it made the observation for the internal position of human body of pathological changes more direct, more distinct. So,the diagnose rate is higher. It has also become essential means and tool. Analysis and processing in medical image are a very important part in medical imageology technology. Using some imagine processing technology, we can process the medical image efficiently and in a deeper level. For example, image denoising, edge detection and segmentation etc., so as to improving the utilization rate for medical image information. It is helpful for drawing information of pathological changes features, and improving accuracy rate of diagnosis.This paper first introduced the medical image processing technology development status, and then introduced the characteristics of the CT images and the basic principles of imaging of CT images, focuses on brain CT image denoising, edge detection and segmentation methods.In the process of denosing for CT images of the brain, because it has some characteristics,such as the biomedical signal is a typical non-stationary signal with a strong noise in background,it is more difficult to resolution. So the paper introduced a method of denoising CT image of the brain based on DEMD (Directional Empirical Mode Decomposition), which can effectively remove noise, then the local features of the signals fully retained. The edge detection of CT image showed weak performances, we use wavelet and the cubic B-spline smoothing filter operator and combined it with the embedded confidence for edge detection. Furthermore, integrating EMD theories,LAWS texture energy measure, and fuzzy clustering ,this thesis proposes CT image segmentation method based on EMD .Finally, the experimental results show that the denoising results clearly when the proposed algorithm compared with other methods, it also be more precise and can be detected in a clear edge images. The segmentation result is ideal.