Understanding and Utilizing Natural Image Statistics

GOAL

One of fundamental problems in computer vision and image processing is the learning of accurate image priors. Most tasks in image processing, computer vision and computational photography involve fundamentally ill-posed inverse problems, solved with the aid of image priors. Despite years of research, current image priors are only approximate, and mostly limited to local patches and kernels. A team of researchers from three universities (Technion, HUJI and Weizmann) attempts to make a fundamental contribution to the learning of more accurate global image priors and to their efficient application to a variety of image restoration problems. More specifically, this team intends to explore the interplay between local patch-based models, and global ones, the role of scale-invariance in such models, and fundamental limits to the performance of various inverse problems using patch-models.

Figure 4: Image Restoration

STATUS

The project started on July 2012
PEOPLE 
Prof. Michael Elad, Technion CS
Prof. Anat Levin, Weizmann Math-CS
Prof. Boaz Nadler, Weizmann Math-CS
Prof. Yair Weiss, HUJI CSE
STUDENTS
Alexander Apartsin
Fredo Durand
Netalee Efrat
William T. Freeman
Daniel Glasner
Tomer Peleg
Yaniv Romano
Daniel Zoran
PUBLICATIONS

Michael Elad ➭

  1. Yaniv Romano, Michael Elad. “Improving K_SVD denoising by post-processing its method-noise”,  ICIP 2013.
  2. Yaniv Romano, Matan Protter, Michael Elad. “Single Image Interpolation via Adaptive Non-Local Sparsity-Based Modeling”, submitted to IEEE-TIP, October 2013.
  3. Tomer Peleg, Michael Elad. A Statistical Prediction Model Based on Sparse Representations for Single Image Super-Resolution”, submitted to IEEE-TIP, November 2013.
  4. Y. Romano and M. Elad, “Improving K-SVD Denoising By Post-Processing Its Method-Noise”, ICIP 2013.
  5. J. Sulam, B. Ophir, and M. Elad, “Image Denoising Through Multi-Scale Learnt Dictionaries”, ICIP 2014.
  6. R. Giryes and M. Elad, “Sparsity Based Poisson Inpainting”, ICIP 2014.

Anat Levin ➭

  1. Netalee Efrat, Daniel Glasner, Alexander Apartsin, Boaz Nadler, Anat Levin, Accurate Blur Models vs. Image Priors in Single Image Super- Resolution”, ICCV 2013.
  2.  I. Gkioulekas, S. Zhao, K. Bala, T. Zickler, Anat Levin, “Inverse Volume Rendering with Material Dictionaries”, SIGGRAPH Asia, ACM Transactions on Graphics, Nov 2013.
  3. A. Levin, D. Glasner, Y. Xiong, F. Durand, W. Freeman, W. Matusik, T. Zickler, “Fabricating BRDFs at High Spatial Resolution Using Wave Optics”. SIGGRAPH, ACM Transactions on Graphics, July 2013.
  4. F. Durand, W. T. Freeman, Boaz Nadler, Anat Levin. “Patch Complexity, Finite Pixel Correlations and Optimal Denoising”. European Conference on Computer Vision (ECCV), Oct 2012.
  5. D. Glasner, T. Zickler and A. Levin “A Reflectance Display”, SIGGRAPH , ACM Transactions on Graphics, Aug 2014

Boaz Nadler ➭

  1. Netalee Efrat, Daniel Glasner, Alexander Apartsin, Boaz Nadler, Anat Levin, Accurate Blur Models vs. Image Priors in Single Image Super- Resolution”, ICCV 2013.
  2.  F. Durand, W. T. Freeman, Boaz Nadler, Anat Levin. “Patch Complexity, Finite Pixel Correlations and Optimal Denoising”. European Conference on Computer Vision (ECCV), Oct 2012.

Yair Weiss ➭

  1. Daniel Zoran, Yair Weiss, Natural Images, Gaussian Mixtures and Dead Leaves”, NIPS 2012, Lake Tahoe, California.
  2. Dan Rosenbaum and Yair Weiss, “Learning the Local Statistics of Optical Flow”. NIPS 2013