DOIONLINE 
DOIONLINE Publish In |
International Journal of Advance Computational Engineering and Networking (IJACEN)-IJACEN |
 Journal Home Volume Issue |
||||||||
Issue |
Volume-9,Issue-9 ( Sep, 2021 ) | |||||||||
Paper Title |
Frame Generative Neural Network for Denoising Severely Noisy Images | |||||||||
Author Name |
Sung Rung Yoo | |||||||||
Affilition |
Texas A&M University, 400 Bizzell St, College Station, Tx 77843, USA | |||||||||
Pages |
25-28 | |||||||||
Abstract |
Object recognition in extreme noisy environments has been an exceedingly difficult task for a long time. The existing denoising technologies still have been challenging either under low PSNR environment or under multiple types of noises. The purpose of this paper is to introduce a new deep learning neural network model named Frame Generative Neural network (FGNN) for denoising extremely noisy images. The FGNN utilizes multiple neural networks connected in parallel to generate many frames. Although each net produces the image with less noise but surely not sufficient, when these frames are synchronously combined, it can produce a much better quality of image with plenty of detail. The experimental results show how effective and promising the FGNN model can be as a tool reducing noise in low PSNR image. Keywords - PSNR, Denoise, Noise, Neural Network, Deep Learning, Frame | |||||||||
| View Paper | ||||||||||
