Pairwise Decomposition with Deep Neural Networks and Multiscale Kernel Subspace Learning for Acoustic Scene Classification.
Erik MarchiDario TonelliXinzhou XuFabien RingevalJun DengStefano SquartiniBjörn W. SchullerPublished in: DCASE (2016)
Keyphrases
- scene classification
- subspace learning
- multiscale
- pairwise
- image representation
- neural network
- canonical correlations
- sparse coding
- visual words
- natural scenes
- semi supervised
- object recognition
- sparse representation
- image classification
- natural images
- dimensionality reduction
- feature space
- biologically inspired
- manifold learning
- pattern recognition
- low dimensional
- kernel function
- data representation
- bag of words
- image processing
- principal component analysis
- similarity measure
- learning algorithm
- kernel methods
- high dimensional data
- face recognition
- scale space
- image features
- high dimensional
- bag of features
- keypoints
- image segmentation
- support vector
- image content
- image patches
- object categories
- metric learning
- affine invariant
- feature extraction