Shape Model´s Alignment Algorithm with Global Shear Transformation

ASM (active shape model) and AAM (active appearance model) model are both parametric model based on statistics. We can locate the key points of a face by AAM model accurately. There is a shape model both in the ASM and AAM. When we build the ASM model and AAM model, we must align the shapes of the training set to a unified framework. In the original alignment process, there are only scale, rotation and translation transformation. In order to improve the accuracy of the alignment, we propose a new algorithm with shear transformation besides the original geometric transformations. At last, the experimental results indicate that the new algorithm is effective for the alignment of the shape model.