Research On Processing CT Medical Images And Implementing DSP-based Hardware

With the development of computer technology and medical imaging technology, computer tomography(CT) is widely applied in the clinical practice. In the recent years, the application of real-time digital image processing technology is very popular in the medical equipment since the continuous progress has been made in digital signal processing technology and dedicated processors.In this thesis, a DSP-based medical CT image processing system is designed with TI Company's TMS3 20DM642 chip as a core.This system could undertake such thing as a CT film reading, and therefore help clinicians to do some diagnosis more accurately and rapidly. A variety of medical CT image processing algorithms, such as the medical image denoising, interpolation, segmentation, registration and edge detection, are systematicly introduced after analysis of the characteristics of the medical CT images at the beginning of the thesis. Still, a new edge detection algorithm based on the fuzzy theory and a new registration algorithm based on an edge extraction and maximum mutual information are proposed , and the system's hardware frame and image processing software frame are designed. Even more, the edge detection algorithm and the registration algorithm proposed in the thesis are transplanted into DSP platform with the aid of TI company CCS programming environment so as to implement real-time medical CT image edge detection function in the medical CT images diagnosis system.This thesis ,finally, points out the direction of future research in this project, that is, along with the further improvement and development of this system, more functions will be achieved, such as real-time denoising, registration, interpolation, segmentation of this medical CT images, and more powerful network data transmission for remote medical care system.