ImVoxelNet: Image to Voxels Projection for Monocular and Multi-View General-Purpose 3D Object Detection.
Danila RukhovichAnna VorontsovaAnton KonushinPublished in: CoRR (2021)
Keyphrases
- multi view
- d objects
- multi views
- multiple views
- single view
- pose estimation
- three dimensional
- viewpoint
- depth map
- single image
- camera calibration
- input image
- depth cues
- image features
- object recognition
- pose parameters
- image sequences
- vanishing points
- semi supervised
- bundle adjustment
- range data
- photometric stereo
- light field
- surface reconstruction
- d mesh
- high resolution
- object detection
- lighting conditions
- reconstruction method
- visual hull
- image synthesis
- surface patches
- camera parameters
- test images
- feature points
- view synthesis
- surface points
- pairwise
- face recognition
- image based rendering
- machine learning