Hybrid Image Denoising with Wavelet and Deep Learning Model

Abstract

Underwater image enhancement restores corrupted underwater photos so that clean images can be produced. Deep learning-based image super resolution reconstruction has beena prominent issue in the near past. To enhance the efficiency and effectiveness of deep learning based solutions, this research puts forth an enhanced picture deep convolutional neural networks-based super-resolution reconstruction technique. Deep learning-based image improvement techniques normally use paired data to train the model; however, in the underwater environment, paired data,such as damaged photographs and their corresponding cleanimages, is difficult to collect simultaneously. Another significant issue is how to effectively keep specific information in the improved image. We provide a neural network-based model for an approach to unpaired underwater picture enhancement to overcome these issues.