Detailed 3D human body reconstruction from multi-view images combining voxel super-resolution and learned implicit representation.
Zhongguo LiMagnus OskarssonAnders HeydenPublished in: Appl. Intell. (2022)
Keyphrases
- human body
- super resolution
- high resolution
- image reconstruction
- three dimensional
- low resolution
- pedestrian detection
- depth map
- multi frame
- pose estimation
- high quality
- single image
- motion estimation
- image representation
- articulated objects
- reconstruction method
- image patches
- depth information
- image processing
- field of view
- multi view
- test images
- image sequences
- surface model