Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders.
Renrui ZhangLiuhui WangYu QiaoPeng GaoHongsheng LiPublished in: CoRR (2022)
Keyphrases
- pre trained
- single image
- input image
- learning algorithm
- image features
- training data
- prior knowledge
- denoising
- image retrieval
- neural network
- image representation
- image segmentation
- probabilistic model
- reinforcement learning
- supervised learning
- small number
- image classification
- feature points
- computer vision
- sparse representation
- keypoints
- image matching
- parametric models
- deep belief networks