Research On Algorithm Of CT/MRI Medical Imaging Segmentation

Image segmentation is one of the important research contents of the medical image processing. It means extracting target objects from medical image accurately. It’s the foundation of3D reconstruction, and the base of medical image processing system in clinical practice. It’s meaningful that realize the purpose fast and accuratly, to satisfy the requirements of the operation plan, to ensure the accuracy of clinical application, and to guide clinicians follow-up work.This paper firstly summaries the basic knowledge of medical image segmentation, including medical image segmentation algorithm, research meaning, research present situation, the segmentation evaluation, etc. Secondly using wavelet decomposition and Markov field to study CT/MRI image segmentation of edge profile information extraction problem, to overcome contour edge positioning is not allowed and the nonstationarity. Continuously, using the robust statistics theory to study the image characteristics, to overcome the image noise instability and using the level set SFM to study the evolution of the image contour.This paper mainly studies as follows:1) A medical image segmentation algorithm based on wavelet transform and Markov random fields is proposed. The algorithm of wavelet multiscale pyramid decomposition get distribution better extract image edge profile information, through the layered Markov modeling, and with the maximum a posteriori probability rule overcome its edge position non-accurate and non-stationary difficult. The experimental results show that the algorithm can improve the quality of image segmentation effectively2) A simple Human-computer interaction segmentation algorithm is proposed. Firstly, Users provided seeds and use robust feature statistics to describe the features of objects, for the reason that it is more suitable for texture and noise images; Secondly, contour evolution fast and accurately according to the sparse field algorithm to get the outline of objects, thus extracting object contour from the MRI and CT image. The results valued by five standards provided by MICCAI show that the segmentation of MRI and CT images of organs and tumors have better results.