| Medical image edge detection is a classical question of image processing andanalysis in virtual surgery system. It plays a crucial role in the virtual surgery system,because it is the foundation in the forward operations, such as medical imagesegmentation, image matching, recognition of focus of disease, and so on.Several traditional and brand new edge detection methods are discussed in detail inthis paper. Their advantages and disadvantages are analyzed objectively. In addition,further study of image edge detection using morphology theory is carried out, whichlay a solid theory foundation for the core task.Due to the existence of such disadvantages as huge computation, long computationtime, and sensitive to noise, a new edge detection method called Omni-directionalmulti-scale multi-structure element is constructed based on the research ofmathematical morphology. Multi-element can extract edges with different shapes,avoiding the missing of edge information using single structure element only.Multi-scale structure element can extract edges with different side, which can extractdetailed information and keep the outline of the image. Omni-directional structureelement can extract the image in different orientations, which can guarantee theintegrity of the image information. Image with noise can be extracted with goodaccuracy and noise-proof by using Omni-directional multi-scale multi-structureelement combined with different weight in different scale. Experiments are carried outon non-noise and noise gray images and the comparison to other edge detectionalgorithm is made to verify the improvement on real-time performance, detectionprecise, and noise proof.The above algorithm is applied to medical image edge detection finally. Theexperimental results show that better image edge could be obtained by our designed morphological algorithm. It has good performance and adaptability, and also keeps thedetailed feature of the original image. |
PDF Full Text Request