Sparse Representations in Image Processing: From Theory to Practice

Description

In this course, you will learn:

• Models are important in data processing, and sparse representation modeling is universal.
• The role of dictionary learning algorithms in applications.
• How to use sparse representations to solve signal and image processing problems.

Syllabus:

1. Overview of the field and this course.
2. Sparseland theoretic and algorithmic background.
3. Introduction to image priors and their evolution in image processing.
4. In-depth view of the Sparseland model including a geometry perspective and processing of Sparseland' signals.
5. Image deblurring and Iterative Shrinkage Thresholding Algorithm (ISTA).
6. Sparesland from an estimation point of view, including a crash-course of estimation theory.
7. The quest for a dictionary: choosing versus learning a dictionary, including basic dictionary learning algorithms: MOD and KSVD.
8. Challenges in dictionary learning and advanced methods, including the double-sparsity, unitary and signature dictionaries.
9. The image denoising problem and ways to solve it, including global and patch-based Sparseland methods.
10. Crash course on SURE estimator for parameter tuning.
11. The tasks of image separation and inpainting, including Morphological Component Analysis (MCA) and global versus patch-based treatment.
12. The single image super-resolution problem and ways to solve it using Sparseland.
13. Course summary and future research directions of the field.