Monocular Differentiable Rendering for Self-Supervised 3D Object Detection.
Deniz BekerHiroharu KatoMihai MorariuTakahiro AndoToru MatsuokaWadim KehlAdrien GaidonPublished in: CoRR (2020)
Keyphrases
- d objects
- pose estimation
- texture mapping
- object recognition
- viewpoint
- multi view
- three dimensional
- cad models
- geometric models
- multiple views
- shape descriptors
- object detection
- computer graphics
- image sequences
- range data
- detection method
- object classes
- d mesh
- image contours
- single view
- physically plausible
- multi views
- three dimensional objects
- human body
- real world objects
- ray tracing
- partial matching
- projective geometry
- optical flow