Data-driven 3D human head reconstruction

Abstract This paper proposes a framework for reconstructing 3D human head models from a single image. Firstly, a preprocessing system for images is designed. The image is automatically segmented using convolutional neural network. The Gabor filter is used to extract the direction of the hair, and a optional manual interaction is used to modify the results. Secondly, the parameters of the FLAME face parametric model are solved by using landmark points as constraint. The high-frequency information of the face image is used to enhance the detail, and the texture map is complemented by the albedo parameterized model. Finally, a strip mesh hair database is constructed, and the hair in the image is reconstructed by using this database and the information extracted from the image.