3D face reconstruction based on non-absolute positive photos and skin model

Reconstruction of a three dimensional (3D) face model is a hot research issue. More and more methods are applied to reconstruct face models. However, there will be produced a certain degree of texture map missing when we use some photos to reconstruct a face model. In this paper, an algorithm based on skin model is proposed to reduce the missing of side texture mapping when we use non-absolute photos to reconstruct a face model. First, we converted RGB color space to YUV and YIQ color space in order to judge which feature points were color points. Secondly, we made a Gamma transforming on the side angle points. Finally, texture mapping after correction according to color points. The three dimensional face model reconstruction processes described in the paper uses some existing theories and apply them to generate the texture map. Experiment results show that this approach can bring about more realistic 3D face models by using side angle photos.