KerCNNs: biologically inspired lateral connections for classification of corrupted images.
Noemi MontobbioLaurent Bonnasse-GahotGiovanna CittiAlessandro SartiPublished in: CoRR (2019)
Keyphrases
- biologically inspired
- image classification
- scene classification
- three dimensional
- input image
- image features
- edge detection
- image analysis
- receptive fields
- image retrieval
- object recognition
- multi modal
- learning rules
- center surround
- biologically plausible
- noise free
- segmentation algorithm
- multiscale
- humanoid robot
- salient features
- computer vision
- colour image segmentation
- image regions
- support vector machine svm
- image representation
- denoising
- support vector machine