Structure+Texture Image Decomposition and its Applications

Decomposing an image into meaningful components is a hot research topic in image processing. It includes image denoising (decomposition of an image corrupted by noise into a signal and noise parts) and structure+texture image decomposition. In particular structure+texture image decomposition has applications in image compression, image restoration, image sharpening, and image segmentation into meaningful parts.

The project consists of two main parts.
1. Analysis, evaluation and comparison of state-of-the-art structure+texture image decomposition methods (for many of them, the authors make their matlab implementations available).
2. Applications of structure+texture image decomposition methods for texture defect detection, object boundary detection, and low-light image enhancement.