2020/04/14 by Enrique Sánchez, Sanchez, Enrique, Michel Valstar +1
Computer Science · Engineering · #3D Shape Modeling and Analysis #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Face recognition and analysis #Generative Adversarial Networks and Image Synthesis
paper · pdf · doi:10.48550/arxiv.2004.07165
openalex publication_date 2020/04/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper addresses a major flaw of the cycle consistency loss when used to\npreserve the input appearance in the face-to-face synthesis domain. In\nparticular, we show that the images generated by a network trained using this\nloss conceal a noise that hinders their use for further tasks. To overcome this\nlimitation, we propose a ''recurrent cycle consistency loss" which for\ndifferent sequences of target attributes minimises the distance between the\noutput images, independent of any intermediate step. We empirically validate\nnot only that our loss enables the re-use of generated images, but that it also\nimproves their quality. In addition, we propose the very first network that\ncovers the task of unconstrained landmark-guided face-to-face synthesis.\nContrary to previous works, our proposed approach enables the transfer of a\nparticular set of input features to a large span of poses and expressions,\nwhereby the target landmarks become the ground-truth points. We then evaluate\nthe consistency of our proposed approach to synthesise faces at the target\nlandmarks. To the best of our knowledge, we are the first to propose a loss to\novercome the limitation of the cycle consistency loss, and the first to propose\nan ''in-the-wild'' landmark guided synthesis approach. Code and models for this\npaper can be found in https://github.com/ESanchezLozano/GANnotation\n