Training Scale-Invariant Neural Networks on the Sphere Can Happen in Three Regimes.
Maxim KodryanEkaterina LobachevaMaksim NakhodnovDmitry P. VetrovPublished in: CoRR (2022)
Keyphrases
- scale invariant
- neural network
- training process
- keypoints
- natural images
- affine invariant
- scale space
- scale invariance
- multiscale
- pattern recognition
- object class recognition
- local binary pattern
- interest point detectors
- translation and scale invariant
- image structure
- texture images
- object recognition
- feature transformation
- rotation invariant
- back propagation
- training set
- training data
- training samples
- higher order
- supervised learning
- image features
- invariant features
- support vector machine
- feature extraction
- image sequences
- scale invariant features
- machine learning