Progressive Face Dynamic Morphing

This work proposes a novel progressive face dynamic morphing approach, where facial structure is well preserved during the morphing process. To this end, we introduce a deep structure aligned model, by which we can achieve continuous alignment between source image and target image with high quality facial structure preservation. Also, a novel semi-supervised 3D expression animation approach is invoked for motion preservation in resulted sequence. Extensive experiments on both MMI facial expression dataset and CK AU-Coded expression dataset %wo common expression datasets demonstrate the effectiveness and efficiency of our model for face dynamic morphing.