Abstract

Image processing has become a crucial component of modern computer vision applications enabling improved visual interpretation automated recognition and advanced machine learning performance. Among the diverse computational methods for enhancing low-quality images the Group Sparse Representation (GSR) algorithm has emerged as a powerful technique for restoring degraded visual data. GSR leverages the principle that natural images exhibit structured sparsity where similar patches within the image can be grouped and represented using shared sparse coefficients. This property allows GSR to effectively handle complex image distortions including blur noise and low illumination. The objective of this research is to analyze the use of the GSR algorithm for simultaneous image enhancement deblurring and noise removal and to evaluate its effectiveness in producing high-quality restoration results compared with conventional algorithms used in image processing. The study explores the workflow of GS