Learning Detailed Radiance Manifolds for High-Fidelity and 3D-Consistent Portrait Synthesis from Monocular Image.
Yu DengBaoyuan WangHeung-Yeung ShumPublished in: CoRR (2022)
Keyphrases
- high fidelity
- image data
- real time
- single image
- learning algorithm
- multiscale
- high resolution
- image features
- image retrieval
- learning process
- image classification
- image content
- image representation
- medical image compression
- lighting conditions
- vector quantization
- online learning
- image segmentation
- multi modal
- pose estimation
- feature points
- multi view
- illumination conditions
- high quality