A novel method of content - based image segmentation using deformable template matching is proposed . a two - dimensional ( 2 - d ) deformable template based on orthogonal curves is built by pre - computing extensions of the deformable template along orthogonal curves and sampling the curves uniformly . then the definitions of internal and external energy functions are given according to the image segmentation problem , and genetic algorithm is used to obtain globally optimal solutions . the proposed method uses a lower - dimensional search space than conventional methods and reduces the sensitivity of the algorithm to initial placement of the template . experiments on real - world images and in simulations at low signal - to - noise ratio show the robustness and good performance of the method 本文提出一种采用可变形模板匹配技术进行基于内容的图像分割算法.通过预先计算出可变形模板沿着变形的正交曲线，并对模板曲线及正交曲线进行离散抽样，建立一基于正交曲线的二维( 2 - d )可变形模板，针对图像分割问题定义控制可变形模板进行变形的内、外部能量函数，本文采用遗传算法搜索能量函数最小的全局最优解.该新算法比传统的可变形模板匹配方法降低了搜索空间的维数，减少了算法对模板初始位置的敏感.对实际图像及模拟低信噪比图像处理的结果表明，新算法具有良好的分割精度及稳定性
The deformable template methods ( dtm ) can resolve some problems that statistical and ann recognition methods can not accomplish . the prominent advantage of dtms is that they can make full use of the prior knowledge on the shape of the characters , and can deal with deformed characters without training work 可变形模板方法可以解决一些统计和神经网络识别方法不能解决的一些问题，其突出的优点在于能充分利用人对字符形状的先验知识，不需要用大量的字符样本进行训练，它对解决小字符集识重庆大学博士学位论文别问题具有一定的什值。
