In All Likelihood, Deep Belief Is Not Enough
The Journal of Machine Learning Research
A probabilistic model for component-based shape synthesis
ACM Transactions on Graphics (TOG) - SIGGRAPH 2012 Conference Proceedings
CoNet: feature generation for multi-view semi-supervised learning with partially observed views
Proceedings of the 21st ACM international conference on Information and knowledge management
Disentangling factors of variation for facial expression recognition
ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part VI
The Shape Boltzmann Machine: A Strong Model of Object Shape
International Journal of Computer Vision
| Hi-index | 0.00 |
The most popular way to use probabilistic models in vision is first to extract some descriptors of small image patches or object parts using well-engineered features, and then to use statistical learning tools to model the dependencies among these features and eventual labels. Learning probabilistic models directly on the raw pixel values has proved to be much more difficult and is typically only used for regularizing discriminative methods. In this work, we use one of the best, pixel-level, generative models of natural images-a gated MRF-as the lowest level of a deep belief network (DBN) that has several hidden layers. We show that the resulting DBN is very good at coping with occlusion when predicting expression categories from face images, and it can produce features that perform comparably to SIFT descriptors for discriminating different types of scene. The generative ability of the model also makes it easy to see what information is captured and what is lost at each level of representation.