Research on Face Image Morphing Based on Automatic Feature Detection

Abstract

Image morphing refers to the technology that can realize the transformation of the source digital image to the target digital image in a smooth transition way. Among them, the most representative is to deform the human face. This paper implements a morphing method based on the Dlib library for automatic detection of facial feature points, including automatic feature detection, warp image and generalized morphing. This method can automatically calibrate the key points of the face more accurately, and use the Delaunay triangulation mesh deformation algorithm to achieve face image morphing. Experiments on multiple pairs of images prove the effectiveness and feasibility of the method.