Demosaicking and Denoising using Deep Learning

Abstract

Demosaicking is the task of recreating full-color image from deficient images framed from overlaid color channel exhibits on image sensors present in digital cameras. And denoising is the process to remove the unwanted noise from the image. Having said that, image demosaicking and denoising is the exceptionally ill-posed problem where at least two third of the image pixels are missing and rest are adulterated by noise. Early methods rely on interpolation techniques for demosaicking but these methods exhibit disturbing visual artifacts. At present convolutional neural networks are the state-of-the-art solution for various image processing tasks. Image demosaicking techniques in the light of convolutional neural networks have demonstrated great results. Despite the fact that there is still space for performance enhancement and edge preservation and good visual quality of the reconstructed image. Therefore, on the basis of aforementioned constraints, the paper presents a novel approach for demosaicking and denoising based on the convolutional neural network to solve this challenge. The propose technique Deep Dense Demosaicking and Denoising using Convolutional Neural Networks (4DCNN) has four phases. In the first phase, the image is organized. In the second phase, the demosaicking is performed using the deep dense convolutional neural network, which give us demosaicked image. In the third phase, denoising performs and pass this image to the final phase. In the fourth phase, and finally image passes to the final post-processing phase producing a higher quality image. To test the feasibility of the proposed scheme, Python language is used. The proposed scheme out performs the several existing methods in terms of throughput delay, latency, accuracy.