Blood Vessels Segmentation Based On CT Images

Image segmentation is a process that the image is divided into multiple regions based on specific needs and characteristics of the image itself. It is the important step of image analysis and pattern recognition and computer diagnosis, also an important area of image processing. In recent years CT images of color developed as a new medical imaging technology, it can provide physicians with more extensive information of patients'body, and it is of great significance in the diagnosis and preoperative preparation of doctors.Nowadays,most of the medical image segmentation is for grayscale images, color medical image segmentation is often extended from the gray image segmentation method In this paper, blood vessels can be extracted form the CT image by an improved watershed algorithm and an algorithm based on genetic algorithms and fuzzy connected image segmentation algorithm.Details are as follows:Because the texture of medical images is complex, the gray contrast is relatively small, the edge is not obvious, this paper combines the algorithm of gray-contrat and wavelet analysis image to enhance the CT images, after this algorithm, the edgeof the images are clearer,and noise reduced. The image enhancement is laid a a good foundation for the next segmentation.For the existence disadvantage of over-segmentation of the watershed algorithm, a automatic segmentation method which can effectively eliminate the local minimum value and the noise of is proposed in this paper. First,comparing the color component gradient images to extract the gradient information.Then, to eliminate invalid gradient information based on multi-threshold segmentation method, finally,finally, introduce the steps of the algorithm and the results. The experimental results show that the gradient image of the approach to conduct a watershed segmentation algorithm can achieve good results.Using the combining algorithm of genetic algorithm and the method of fuzzy connection to segment CT images.First, select the RGB color space.Then the path strength as the fitness function of genetic algorithm to find the optimal path between the seed point and the other points, the strength of the path is fuzzy connectedness between the point and the seed points Finally, all the fuzzy connectedness as gray values, the objective and overlap the background image can be shown, the object can be extracted by the seted threshold.The experimental results show that the improved watershed segmentation algorithm achieved a good result,comparing with the conventional watershed algorithm there are less basins, greatly improved the watershed segmentation algorithm. Comparing with the method of fuzzy connection, small blood vessels can also be split out to a good segmentation performance by the combining algorithm of genetic algorithm and the method of fuzzy connection. This can lay a good foundation for Doctors'diagnosis and the preoperative preparation of operation.