基于加权主动形状模型人脸定位研究

Research of Location of Human Face Based on Weighted ASM

主动形状模型方法常常用于人脸图像描述,但由于受到初始情况、光照等诸多因素的影响,主动形状模型容易陷入最优化过程中的最小问题,为了解决此问题,用一种基于评价信息的加权主动形状模型(Weighted-ASM)或者WASM[1],首先描述局部纹理模型,然后用了一种形状评价函数衡量描述得到的形状与训练数据的匹配程度。WASM采用形状评价信息,把搜索得到的形状用加权的方式投影到形状子空间,加权投影可以利用搜索过程中的信息,使得搜索可能跳出局部极值,从而得到更准确结果。实验证明该方法非常有效。

Active Shape model described human face images,but the result of ASM was often influenced by some factors such as the initial location.,illumination and so on,which led to the local minima in optimization. In this paper,a shape evaluation method called weighted ASM or WASM was proposed. Firstly,weighted ASM described the local appearance model of ASM. Secondly,it proposed the shape evaluation function which determined how well the searching shape match models derived from the training set. The weighted-ASM used this evaluation information to project the searching shape into the shape space in a weighted way. The weighted projection can be used in the process of searching information,the weighted projection can drag the search out of local minima to get more accurate results. Experiment results demonstrate the efficacy of the method in locating facial features.