Unsupervised Learning of Depth and Depth-of-Field Effect from Natural Images with Aperture Rendering Generative Adversarial Networks.
Takuhiro KanekoPublished in: CoRR (2021)
Keyphrases
- natural images
- unsupervised learning
- natural scenes
- sparse coding
- object recognition
- higher order
- image denoising
- denoising
- multiscale
- input image
- image patches
- natural image statistics
- depth map
- compressed sensing
- test images
- receptive fields
- power spectrum
- image statistics
- high quality
- prior models
- statistics of natural images
- low level vision tasks
- scale invariant
- learning algorithm
- image quality assessment
- computer graphics
- generative model
- semi supervised
- machine learning