A truncated EM approach for spike-and-slab sparse coding.
Abdul-Saboor SheikhJacquelyn A. SheltonJörg LückePublished in: J. Mach. Learn. Res. (2014)
Keyphrases
- sparse coding
- generative model
- unsupervised learning
- dictionary learning
- natural images
- sparse representation
- expectation maximization
- linear combination
- image classification
- single image super resolution
- em algorithm
- image representation
- discriminative power
- feature space
- person re identification
- mixture model
- maximum likelihood
- probabilistic model
- group sparsity
- image patches
- supervised learning
- multiscale
- sparse codes
- gaussian mixture model
- semi supervised
- high dimensional
- visual words
- machine learning
- neural network
- prior knowledge
- k means
- sparsity constraints
- object recognition
- model selection
- dimensionality reduction
- denoising
- object detection
- image segmentation
- object category recognition