Locally adaptive bilateral clustering for image deblurring and sharpness enhancement

This paper presents an integrated deblurring and sharpness enhancement algorithm for restoration of images corrupted by mild blur and improvement of image details. The proposed algorithm is capable of capturing local image structure by adaptively clustering pixels geometric and radiometric features based on their local augmented variational series. By selectively utilizing the segmented features, deblurring and/or sharpening can be effectively performed without either producing noise magnification or over-sharpening artifacts. Experimental results from both simulated and real image data, which are comparable to recently published results, show that the proposed algorithm is superior over the state-of-the-arts. In addition, the proposed method is fully automated with no manual set-up and it enjoys the luxury of fast processing time. These make the proposed algorithm practically useful for image applications in consumer electronic products.