Adaptive Radii selection based Inpainting method for impulse noise removal

Digital images acquired via electronic products are prone to corruption by various noises among which Salt & Pepper noise is one such variant. In cases of corruption by said noise, it is important to apply image restoration techniques that regenerate corrupted pixel values. In this document, an Adaptive Radii Selection based Inpainting Method is proposed, which makes use of the Telea Inpainting algorithm by dynamically choosing the optimal radii with which the best denoised pixel can be obtained. The radii selection is based on observed noise density in a particular window around the noisy pixels. Further, a recombination logic is applied to the inpainted images to obtaining the restored image. Qualitative and Quantitative analysis have been performed on standard images as well as the Kodak image dataset. This work serves as an extension to the Telea Inpainting algorithm for denoising efforts showing an improvement in quality metrics such as PSNR by 0.31dB and 8-12% in IEF.