Image noise removal framework based on morphological component analysis

Abstract

Now image denoising is an important process in image processing. The proposed method focuses on rain streak removal frame work based on morphological component analysis. Bilateral filter is used in the denoising stage. Then the filtered image partitioned into low frequency and high frequency component. The high frequency component undergone various processes such as patch extraction, dictionary learning and dictionary partitioning. The output of dictionary partitioning approach undergone morphological component analysis as an image decomposition process. As a result, the rain component can be successfully removed from the image while preserving most of the original image details.