Texture Modeling with Convolutional Spike-and-Slab RBMs and Deep Extensions

Heng Luo, Pierre Luc Carrier, Aaron Courville, Yoshua Bengio
Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, PMLR 31:415-423, 2013.

Abstract

We apply the spike-and-slab Restricted Boltzmann Machine (ssRBM) to texture modeling. The ssRBM with tiled-convolution weight sharing (TssRBM) achieves or surpasses the state-of-the-art on texture synthesis and inpainting by parametric models. We also develop a novel RBM model with a spike-and-slab visible layer and binary variables in the hidden layer. This model is designed to be stacked on top of the ssRBM. We show the resulting deep belief network (DBN) is a powerful generative model that improves on single-layer models and is capable of modeling not only single high-resolution and challenging textures but also multiple textures with fixed-size filters in the bottom layer.

Cite this Paper

BibTeX
@InProceedings{pmlr-v31-luo13a, title = {Texture Modeling with Convolutional Spike-and-Slab RBMs and Deep Extensions}, author = {Luo, Heng and Carrier, Pierre Luc and Courville, Aaron and Bengio, Yoshua}, booktitle = {Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics}, pages = {415--423}, year = {2013}, editor = {Carvalho, Carlos M. and Ravikumar, Pradeep}, volume = {31}, series = {Proceedings of Machine Learning Research}, address = {Scottsdale, Arizona, USA}, month = {29 Apr--01 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v31/luo13a.pdf}, url = {https://proceedings.mlr.press/v31/luo13a.html}, abstract = {We apply the spike-and-slab Restricted Boltzmann Machine (ssRBM) to texture modeling. The ssRBM with tiled-convolution weight sharing (TssRBM) achieves or surpasses the state-of-the-art on texture synthesis and inpainting by parametric models. We also develop a novel RBM model with a spike-and-slab visible layer and binary variables in the hidden layer. This model is designed to be stacked on top of the ssRBM. We show the resulting deep belief network (DBN) is a powerful generative model that improves on single-layer models and is capable of modeling not only single high-resolution and challenging textures but also multiple textures with fixed-size filters in the bottom layer.} }
Endnote
%0 Conference Paper %T Texture Modeling with Convolutional Spike-and-Slab RBMs and Deep Extensions %A Heng Luo %A Pierre Luc Carrier %A Aaron Courville %A Yoshua Bengio %B Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2013 %E Carlos M. Carvalho %E Pradeep Ravikumar %F pmlr-v31-luo13a %I PMLR %P 415--423 %U https://proceedings.mlr.press/v31/luo13a.html %V 31 %X We apply the spike-and-slab Restricted Boltzmann Machine (ssRBM) to texture modeling. The ssRBM with tiled-convolution weight sharing (TssRBM) achieves or surpasses the state-of-the-art on texture synthesis and inpainting by parametric models. We also develop a novel RBM model with a spike-and-slab visible layer and binary variables in the hidden layer. This model is designed to be stacked on top of the ssRBM. We show the resulting deep belief network (DBN) is a powerful generative model that improves on single-layer models and is capable of modeling not only single high-resolution and challenging textures but also multiple textures with fixed-size filters in the bottom layer.
RIS
TY - CPAPER TI - Texture Modeling with Convolutional Spike-and-Slab RBMs and Deep Extensions AU - Heng Luo AU - Pierre Luc Carrier AU - Aaron Courville AU - Yoshua Bengio BT - Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics DA - 2013/04/29 ED - Carlos M. Carvalho ED - Pradeep Ravikumar ID - pmlr-v31-luo13a PB - PMLR DP - Proceedings of Machine Learning Research VL - 31 SP - 415 EP - 423 L1 - http://proceedings.mlr.press/v31/luo13a.pdf UR - https://proceedings.mlr.press/v31/luo13a.html AB - We apply the spike-and-slab Restricted Boltzmann Machine (ssRBM) to texture modeling. The ssRBM with tiled-convolution weight sharing (TssRBM) achieves or surpasses the state-of-the-art on texture synthesis and inpainting by parametric models. We also develop a novel RBM model with a spike-and-slab visible layer and binary variables in the hidden layer. This model is designed to be stacked on top of the ssRBM. We show the resulting deep belief network (DBN) is a powerful generative model that improves on single-layer models and is capable of modeling not only single high-resolution and challenging textures but also multiple textures with fixed-size filters in the bottom layer. ER -
APA
Luo, H., Carrier, P.L., Courville, A. & Bengio, Y.. (2013). Texture Modeling with Convolutional Spike-and-Slab RBMs and Deep Extensions. Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 31:415-423 Available from https://proceedings.mlr.press/v31/luo13a.html.

Related Material