Self-Supervised Pre-training for 3D Point Clouds via View-Specific Point-to-Image Translation.
Qijian ZhangJunhui HouPublished in: CoRR (2022)
Keyphrases
- image analysis
- image data
- image features
- single image
- image classification
- image representation
- multiscale
- template matching
- input image
- high resolution
- virtual camera
- image content
- low level
- image retrieval
- image pixels
- point cloud
- region of interest
- test images
- hough transform
- image regions
- training set
- image matching
- image set
- pixel values
- edge detection
- structure from motion
- multiresolution
- segmentation method
- object recognition
- similarity measure
- gray level
- image collections