Centerline Extraction Of Vascular Based On Minimal Cost Path And Wavelet

In the field of medical visualization, effective extraction of centerline has been one study and exploration direction to researchers. To extract a better centerline of tubular structure in image not only make virtual path navigation more accurate to medical diagnosis, but also make feature extraction more convenient so as to further research.To some extent, although use minimal cost path method to extract centerline in a particular form of image data can also achieve the desired effect, but in the details of the branches, this centerline can?t properly be extracted .At the same time, in complexity there are also shortcomings.This article, based on image property and feature, most important that it also based on current research analysis of medical image centerline extraction, suggest the more fitness, more accurate centerline extraction method which using wavelet techniques refined method to extract centerline, not only in the details of the branches centerline can be very good extracted, but also it more efficient because of the extraction process be limited in center region of vascular.The shortest path, which uses the minimum path method, can extract the vascular centerline. Although this path has the shortest costs but it also has the uncertainty, that is, it is likely to get the minimal path but it is not the target, which is an ideal centerline.In this paper, wavelet technology is used to extract central region, first, we use this technology to the tubular structure in order to reduce its central region, while the minimum path be extracted within this extracted central region. This method is not like other methods which in the overall image to extract the centerline, but in the centered region to find the minimum path. So, use our method to extract centerline of vascular, not only in computational complexity, but also in accuracy can achieve very good results.By the comparing with other proposed methods, our proposed method in this paper can be proved better than others. Not only better in the calculation of performance than other methods but also better in the accuracy of the centerline extraction.This paper proposed extraction method for medical diagnosis or application of visualization can be done better. Both in theory and in application sides, there are great significance and practical reference value.