Two-dimensional Edge Detection Of Medical Images And Three-dimensional Triangulation

The system of medical image processing and analysis is one of popular research projects; it is an interdisciplinary research as well as an important application of computer graphics and image process in the biomedical engineering. The segmentation and collection of medical image are still the focus of the research, with great prospect in the medical image processing, which deal with much relevant knowledge referring to image process, computer graphics, and medical field.The main contents of my thesis are divided into three parts including: 1.edge detection of the medical image; 2.tumor segmentation through 8-connected labeling and human-computer Interactive approach; 3. reconstruction and 3D segmentation based on ray casting.The main work and innovations of this thesis manifest themselves in following four aspects:1. Based on the problem of the edge detection, we describe the edge types existing in real images in terms of mathematical models. Then we classify present edge detection operators and introduce the advantages and disadvantages of different edge detection operators.2. Through improving the traditional canny algorithm by implementing double thresholds and angle ambiguity algorithm, on the one hand we improve the accuracy of edge extraction; on the other hand we speed up the edge extraction. At last, we extract the edge of tumor by combining 8-connected labeling and human-computer Interactive approach.3. In order to turn out the process converting 2D to 3D reconstruction, we reconstruct 2D medical image implemented in ray-tracing volume rendering algorithm, which can get a 3D model. 4. We have to make segmentation on 3D model by algorithm of space plane and cube segmentation in order to observe the segmented tissues and organs from the three-dimensional perspective.