1 Introduction In this paper,we propose a novel domain-adaptive reconstruction method that effectively leverages deep learning and synthetic data to achieve robust 3D face reconstruction from a single depth image.The ...1 Introduction In this paper,we propose a novel domain-adaptive reconstruction method that effectively leverages deep learning and synthetic data to achieve robust 3D face reconstruction from a single depth image.The method applies two domain-adaptive neural networks for predicting head pose and facial shape,respectively.Both networks undergo training with a customized domain adaptation strategy,using a combination of auto-labeled synthetic and unlabeled real data.展开更多
文摘1 Introduction In this paper,we propose a novel domain-adaptive reconstruction method that effectively leverages deep learning and synthetic data to achieve robust 3D face reconstruction from a single depth image.The method applies two domain-adaptive neural networks for predicting head pose and facial shape,respectively.Both networks undergo training with a customized domain adaptation strategy,using a combination of auto-labeled synthetic and unlabeled real data.
