Mixtures of conditional Gaussian scale mixtures applied to multiscale image representations
Preprint
Abstract
We present a probabilistic model for natural images which is based on Gaussian scale mixtures and a simple multiscale representation. In contrast to the dominant approach to modeling whole images focusing on Markov random fields, we formulate our model in terms of a directed graphical model. We show that it is able to generate images with interesting higher-order correlations when trained on natural images or samples from an occlusion based model. More importantly, the directed model enables us to perform a principled evaluation. While it is easy to generate visually appealing images, we demonstrate that our model also yields the best performance reported to date when evaluated with respect to the cross-entropy rate, a measure tightly linked to the average log-likelihood.Keywords
All Related Versions
- Version 1, 2011-09-20, ArXiv
- Published version: PLOS ONE, 7 (7), e39857.
This publication has 6 references indexed in Scilit:
- Lower bounds on the redundancy of natural imagesVision Research, 2010
- Neural Mechanisms of Natural Scene Perception2008
- From information scaling of natural images to regimes of statistical modelsQuarterly of Applied Mathematics, 2007
- Occlusion Models for Natural Images: A Statistical Study of a Scale-Invariant Dead Leaves ModelInternational Journal of Computer Vision, 2001
- Independent component filters of natural images compared with simple cells in primary visual cortexProceedings Of The Royal Society B-Biological Sciences, 1998
- Adaptive Mixtures of Local ExpertsNeural Computation, 1991