| Conference/Proceedings | 2018 IEEE International Conference on Multimedia Expo Workshops (ICMEW) |
| Start date | 23.07.2018 |
| End date | 27.07.2018 |
| Address | San Diego, CA, USA |
| Organisation | IEEE |
| Author(s) | Rolf Jongebloed, Ruben Verhack, Lieven Lange, Thomas Sikora |
| Title | Hierarchical Learning of Sparse Image Representations using Steered Mixture-of-Experts |
| Abstract | Previous research showed highly efficient compression results for low bit-rates using Steered Mixture-of-Experts (SMoE), higher rates still pose a challenge due to the non- convex optimization problem that becomes more difficult when increasing the number of components. Therefore, a novel estimation method based on Hidden Markov Random Fields is introduced taking spatial dependencies of neighbor- ing pixels into account combined with a tree-structured split- ting strategy. Experimental evaluations for images show that our approach outperforms state-of-the-art techniques using only one robust parameter set. For video and light field mod- eling even more gain can be expected. |
| Key words | compression, Bayes methods, mixture models, Bayesian model, nonstationary random process, piecewise stationary processes, pixel values, space-continuous Gaussian mixture model, statistical representation, steered mixture-of-experts, Analytical models, Estimation, Gaussian mixture model, Image reconstruction, mixture models, mixture-of-experts, Hierarchical Learning |
| File | 1536Jongebloed2018.pdf |