Name
Abstract | ||
|---|---|---|
To enhance the understanding of human perception and mimic it into an artificial system, several types of graphical models have been proposed that emulate the functionality of neurons in biological neural networks. In this work, we investigate the discriminatory power of two such probabilistic models of vision: a multivariate Gaussian model [1] and a restricted Boltzmann machine [2], both widely used to solve classification problems in computer vision. We quantify the generative ability of these models on standard benchmark data sets and show that neither approach on their own is powerful enough to carry out vision tasks because of the very low discrimination they achieve. There is clearly a need for inducing discrimination by a mechanism that exploits these generative models. We show that the Fisher kernels [3] derived from both the Gaussian and restricted Boltzmann machine can significantly improve the classification performance on benchmark tasks while maintaining the biological plausibility of its implementation [4]. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.1109/MLSP.2013.6661977 | Machine Learning for Signal Processing |
Keywords | Field | DocType |
Boltzmann machines,Gaussian processes,computer vision,graph theory,image classification,image recognition,probability,artificial system,benchmark data sets,biological neural networks,biologically inspired models,computer vision,fisher kernels,generative models,graphical models,human perception,image classification problems,multivariate Gaussian model,neuron functionality,probabilistic models,restricted Boltzmann machine,visual scene recognition,Deep Learning,Fisher Kernel,Multivariate Gaussian model,Restricted Boltzmann Machine | Restricted Boltzmann machine,Biological plausibility,Boltzmann machine,Pattern recognition,Computer science,Gaussian process,Artificial intelligence,Probabilistic logic,Graphical model,Artificial neural network,Contextual image classification,Machine learning | Conference |
ISSN | Citations | PageRank |
1551-2541 | 1 | 0.36 |
References | Authors | |
14 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Tayyaba Azim | 1 | 31 | 3.89 |
| Mahesan Niranjan | 2 | 775 | 120.43 |