ScanGAN360: A Generative Model of Realistic Scanpaths for 360$^{\circ}$ Images.
Daniel MartinAna SerranoAlexander W. BergmanGordon WetzsteinBelén MasiáPublished in: CoRR (2021)
Keyphrases
- generative model
- probabilistic model
- bayesian framework
- visual patterns
- image data
- input image
- image features
- object parts
- image analysis
- illumination conditions
- em algorithm
- human hair
- feature representations
- discriminative learning
- object categories
- test images
- image classification
- image retrieval
- object recognition
- eye tracking
- topic models
- lighting conditions
- semi supervised
- active learning
- prior knowledge
- decision trees
- computer vision