Transferring scale-independent features to support multi-scale object recognition with deep convolutional neural network.
Xiran ZhouPublished in: SIGSPATIAL/GIS (2018)
Keyphrases
- convolutional neural network
- object recognition
- multiscale
- image features
- scale space
- scale invariant
- object class recognition
- feature detection
- visual object recognition
- feature set
- natural images
- keypoints
- end users
- multiple scales
- image processing
- single scale
- automatic scale selection
- invariant features
- image structure
- face detection
- image representation
- classification accuracy
- low level
- feature vectors
- detection algorithm
- image classification
- edge detection
- multi class
- recognizing objects
- coarse scale
- image retrieval
- feature space