Convolutional Neural Networks for Face Illumination Transfer

Face images contain rich and diverse information and have become the focus of research in the field of computer vision. The current research on facial images includes facial recognition, facial image makeup transfer, facial image segmentation, facial image illumination transfer, facial image beautification and rendering. Among them, the illumination of face images is the focus of research. This paper proposes and implements a illumination transfer method combined with a deep neural network to obtain a illumination transfer result that is closer to the real illumination effect. This method mainly implements illumination transfer through illumination model training and illumination classification, illumination matching, and based on the transfer of image style transfer. First, by combining the convolutional neural network to complete the classification of illumination on the face dataset, a model that can classify the illumination of the face image is obtained; then use the model to achieve the illumination matching of a single face image and obtain an image similar to the illumination of a given face image from the face image illumination dataset; finally, through the illumination classification model, the given reference face illumination image is extracted and processed for the relevant illumination features to facilitate the transfer to the input face image so that realizes the illumination transfer to the entire face image, including the neck, etc., and can move the illumination in multiple directions.