Analysis of Image Quality and Video Denoising Using Convolutional Neural Networks

Capturing digital images and videos has increased invariably in recent times due to the advancement of cameras in smartphones. The images are generally found to be deteriorated in terms of quality due to noise interference from various sources or because of clicking during motion. The proposed work concentrates on the denoising techniques of an image with the help of UNET with various filtering techniques. The structure of the CNN is exploited for removing the noise in the image by passing it through various layers for up-sampling and down-sampling, in addition to preserving the spatial information. The idea of noise removal in the images is extended to videos as an application of our main research area and the optimal results are obtained using the same algorithm trained for image denoising. A comparative analysis of a various deep learning denoising algorithms is performed in terms of image performance parameters like Peak-Signal-to-Noise-Ratio (PSNR) and Structural Similarity Index (SSIM).