Recurrent Connections Aid Occluded Object Recognition by Discounting Occluders.
Markus Roland ErnstJochen TrieschThomas BurwickPublished in: ICANN (3) (2019)
Keyphrases
- object recognition
- d objects
- computer vision
- object tracking
- partially occluded
- image understanding
- object detection
- image representation
- object class
- feature matching
- template matching
- object categorization
- object categories
- recurrent neural networks
- image features
- feed forward
- image matching
- natural images
- scene understanding
- scene classification
- shape recognition
- constrained search
- multi unit combinatorial auctions
- graph matching
- spiking neural networks
- recurrent networks
- recognition scheme
- object classification
- database
- unsupervised learning
- hidden markov models
- genetic algorithm