| Oral disease incidence is increasing. Improving the accuracy of detection equipment is an urgent need to help dental doctor make an accurate diagnosis of the disease and achieve a better therapeutic effect. Clear image segmentation will contain abundant information about organs that help doctors get the situation of the lesion site and improve the therapeutic effect. These goals all require clearer definition of image segmentation. FCM image segmentation algorithm has overcomed the lack of the traditional hard division, because it avoids the missing deficiency of pixels information. With the help of soft partitioning method, FCM algorithm will deal with the pixels include the adjacent information of pixels, which can improve the accuracy of image segmentation. In this paper, according to the recent research, we summarize the hotspots of current dental equipment demands and put forward the WIPFCM algorithm(Weighted Image Patch-based FCM algorithm), which is built based on the FCM weighted image patch segmentation algorithm. In this algorithm, the image patches will instead of the pixels in fuzzy clustering process and construct a weighting scheme,this can make the pixels of each image patch anisotropic weights. Therefore, the proposed algorithm will take the spatial information into account in the image segmentation process. By this method, the robustness of noise will be improved and get better image segmentation effect at the same time. By comparing the improved fuzzy c-means algorithm and several other segmentation methods, experiment results show that the proposed WIPFCM can effectively overcome the noise effect, significantly improve the clarity and accuracy of image segmentation and retain more image details. By the oral images and synthetic images segmentation, the segmentation effect of the proposed algorithm WIPFCM obviously is better than other algorithms which are based on the FCM, the noise also has a higher robustness. This paper focuses on FCM clustering image segmentation technology and how to improve its performance.First:The work starts from the research purpose and the significance of the research. We will understand the domestic and international research on the image segmentation algorithm, and compare the current development status of oral CT. The overall framework of research on segmentation algorithm of low dose oral CT image will be made.Second:Studying the basic theories of the system. the basic knowledge in this paper is:the theory of traditional oral CT, the cone beam CT technology, the concept of fuzzy set, the fuzzy clustering method based on FCM and image quality evaluation method. We will analyze its working principle and do a comparative experiment about the traditional methods to improve its performance.Third:The third part is simulating experiments about FCM algorithm in the actual application. We compare the segmentation effect of FCM algorithm with the improved FCM algorithm, introduce the FCM algorithm validity evaluation index, analyze the deficiencies of the existing methods and its future improvement direction.Fourth:In this part, the proposed algorithm will be implemented to verify the experimental results. The WIPFCM algorithm doesn’t add a constraint to the target function. Instead, in the image segmentation process, image patches will instead the pixels, in this way, the spatial information will be considered directly. The experiment results show this method can protect the image detail better and resist the influence of noise. Through the segmentation of coral images and synthetic images, the experiment results show the proposed WIPFCM algorithm can achieve better image segmentation. |
PDF Full Text Request