A Case for Object Compositionality in Deep Generative Models of Images.
Sjoerd van SteenkisteKarol KurachSylvain GellyPublished in: CoRR (2018)
Keyphrases
- generative model
- object parts
- visual patterns
- input image
- probabilistic model
- discriminative learning
- image features
- keypoints
- multiple objects
- image set
- image retrieval
- partial occlusion
- image classification
- object recognition
- prior knowledge
- object class
- bounding box
- conditional random fields
- test images
- target object
- object categories
- computer vision
- object localization
- image regions
- natural images
- semi supervised
- discriminative methods
- discriminative models
- articulated objects
- object detectors
- feature points
- bayesian networks
- similarity measure