基于PET医学图像边缘检测算法的研究

目的:边缘检测在图像处理中至关重要,可被广泛应用于目标区域识别、区域形状检测、图像分割等图像分析领域。边缘是图像中不平稳现象和不规则结构的重要表现,往往携带着图像中的大量信息,并给出图像轮廓。在医学图像三维显示技术中,为了更精确的临床判别需要得到单像素的清晰轮廓,因此我们提出一种新的边缘检测算法。方法:在传统的小波边缘检测的基础上,提出了一种新的边缘算法,即基于小波极大值边缘检测算法,应用模糊算法构造相应的隶属函数,再对得到的极大值进一步筛选。结果:将该算法应用到医学图像中,最终可以得到较清楚的单像素边缘轮廓,实验结果证明了该算法的可行性。结论:运用这种算法处理过的医学图像边缘锐化更好,更清晰,能够为肿瘤的早期识别提供依据,满足医学影像识别的需要。

The Edge Detection Research Based on PET Medical Images*

Objective： The edge detection of digital image is a very important foundation of the image analysis field, such as target area identification, regional shape detection, image segmentation and so on. The edge is the important performance between the rough phenomenon and the irregular structure of the image. It often carries most of the image information and gives the outline of the image. In the three dimensional display technology of medical image, hospital need get the more accurate clinical criterion,so we need a new edge algorithm. Methods： This paper put forward a new edge algorithm which is based on the traditional wavelet edge detection. The algorithm is based on wavelet maximum edge detection algorithm. First we construct a corresponding subordinate function by fuzzy algorithm; then we filter the maximum values of the obtained. Results： Applied this algorithm to medical image, finally can get a clear single pixel edge profile and the results of the experiment prove the feasibility of the algorithm. Conclusion： The medical image edge treated by this algorithm seems be more better and clearer. It can provide the basis for the the early recognition of tumor and meet the needs of the medical image recognition.