Video Denoising Using Convolutional and Deep Neural Networks

Removal of noise from video files has become as one of the most important academic worries and especially with the growing reliance on video data in domains such as security, autonomous vehicles, and entertainment. Recent improvements in deep learning have lead to powerful neural network-based methods that go over traditional Denoising techniques. In this research, a number of deep learning algorithms were tested for video reduced noise on an array of video files. The results demonstrated the effectiveness of deep learning models in reducing noise artifacts, by on the mean square error (MSE) metric, which measures the difference between the original clean videos and the Denoise outputs. The research included processing (4) video files Which includes a number of videos of different lengths, and the number of frames in this work has become 1,604.through both (convolutional neural network ((CNN)) and (deep neural network (DNN)) and the results showed the superiority of the (CNN) method over the (DNN) method through the results of the (4) deference experiment. The best experiment with minimum MSE was (, μ=0, =0.02) with (MSE DNN =0.02001410), the methods can be applied to other multimedia files (audio, image).