3D-LMNet: Latent Embedding Matching for Accurate and Diverse 3D Point Cloud Reconstruction from a Single Image.
Priyanka MandikalNavaneet K. L.Mayank AgarwalVenkatesh Babu RadhakrishnanPublished in: BMVC (2018)
Keyphrases
- single image
- shape recovery
- d scene
- shape from shading
- image pairs
- multiple images
- indoor scenes
- light source
- matching algorithm
- structured light
- partially occluded
- image correspondences
- three dimensional
- super resolution
- feature matching
- high resolution
- latent variables
- intrinsic images
- reconstruction method
- image matching
- image reconstruction
- stereo correspondence
- scene geometry
- high quality
- feature extraction