多灰度图像连续形变的计算机识别技术研究

Research on Techniques of Recognition of Continuous Deformation of Images with Multi-Gray-Levels

采用网格分层细化技术并且用对网格的处理代替已有的逐点处理,把连续形变过程分为拓扑保持性连续形变和拓扑非保持性连续形变.文中算法可以任意调节处理精度和处理时间,从而满足不同层次、不同场合的应用需要.实验结果证明了该算法的良好性能.

In this paper, we adopt a skill of remeshing procedure and an idea in classification of continuous deformation into those with and without preserving topological structures, and introduction of a new method for identifying continuous deformation. The main advantages of this algorithm are the flexible modulations between the accuracy to calculate and the time to process, and so the needs from different applications could be satisfied easily. At last, a few examples are given to test the versatility of the techniques, and it shows that the algorithm developed here exhibits good performance.