A Hybrid Approach of Adam Optimization with CNN to Remove Noise on Images
Manjunath Prasad, N. P. Subiramaniyam · 2022 International Mobile and Embedded Technology Conference (MECON) · 2022
A digital picture was tainted by noise as a result of the surroundings. There is a novel convolutional neural network for eliminating pictures. The suggested solution incorporates planning approaches with a convolutional neural network. a superior normalizing procedure, the digital picture is initially subjected to the Batch preprocessing step. The picture would then be fed into a Convolution Neural Network (CNN) with a Leaky rectified linear component, which extracts the feature. The CNN algorithm reconfigures the properties of the image that has been retrieved. Ultimately, the rectified picture was integrated Adam optimization technique and Deep CNN with MSE error function to educate and optimize the extracted features. The suggested platform’s performance was evaluated at different sound levels as well as contrasted to different pads and steps. The suggested Deep CNN with Adam optimization method surpasses earlier denoising techniques and gives a satisfactory performance in terms of peak signal-to-noise ratio (PSNR).