Multifractal-based Image Segmentation Method With Application In Medical Image

Image segmentation is the basic technology in image process and computer vision fields,and also is the important part of image analysis and computer vision system.Medical image segmentation is the main applied domain in image segmentation. It is more and more important during clinical diagnosis and treatment.Multifractal is to make a description of a measure over a region both locally and globally.First , this thesis give the method of calculating the singularity exponents, and calculate the multifractal spectrum based on box-counting. Then we could generate the spectrum images and describe how the spectrum can be used to perform image segmentation and texture classification.Second, according the multifractal theory, this thesis tested using the CT Image of the brain. This thesis analyze the predominance of the multifractal arithmetic compare with the arithmetic operators, and get different edge information from different value of the spectrum.Finally, this thesis use the spectrum of CT image with region growing and region merging to segment the CT image. And design the medical image segmentation system use the TI's digital signal processor. Then display the clear result image in the screen.The results of the simulative experiment on PC and the performance of the DSP evaluation both testify the feasibility and validity of the Methods of medica1image segmentation based on multifractal by this paper, and it has good practical value in application.