Classification of Natural Images Using Supervised and Unsupervised Classifier Combinations.
Leena LepistöIivari KunttuJorma AutioAri VisaPublished in: ICIAP (2005)
Keyphrases
- natural images
- supervised classification
- supervised learning
- unsupervised learning
- supervised training
- natural scenes
- feature selection
- multiscale
- image denoising
- class labels
- object recognition
- feature space
- decision trees
- denoising
- higher order
- sparse coding
- support vector machine
- unsupervised methods
- superpixels
- training data
- semi supervised
- support vector
- image patches
- input image
- power spectrum
- training set
- receptive fields
- weakly supervised
- scale invariant
- feature extraction
- machine learning
- restricted boltzmann machine
- text classification
- image statistics
- statistics of natural images
- learning algorithm
- image classification
- feature set
- natural image statistics
- support vector machine svm
- model selection
- classification accuracy
- active learning
- test images
- image processing
- feature vectors
- deep learning
- image representation
- visual cortex
- markov random field
- prior models
- saliency map
- computer vision
- back propagation
- conditional random fields
- neural network