Application of Mathematical Modelling and Deep Learning in Image Restoration using Edge Preservation Method

Abstract

The image restoration has witnessed significant advancements with the integration of deep learning techniques with offering unparalleled capabilities in enhancing image quality. This paper study proposed a novel approach to image restoration by incorporating a sophisticated edge preservation method using the deep learning framework. The method aims to address the challenge of preserving high-frequency details, such as edges, while restoring images from various forms of degradation. Also we investigate the use of deep neural networks trained on a different noisy image to restore a clean image with preserving edges of original image. The Deep convolutional neural network (DCNNs), and ResNet50 in deep learning model learns elaborate patterns and features, enabling the reconstruction of images with improved clarity and fidelity. The proposed Deep Convolution Neural Network and ResNet50 methods are designed to restore image content with intelligently preserve and enhance edge information, crucial for maintaining the structural integrity of the original scene. The proposed model is trained on diverse datasets, encompassing a wide range of image degradations, ensuring robust performance across various real-world scenarios. Experimental results demonstrate the efficacy of the proposed approach in comparison to existing methods, showcasing superior edge preservation and overall restoration quality. This research contributes to the advancement of image processing techniques, offering a powerful tool for applications such as medical imaging, satellite imagery, and digital photography, where maintaining fine details is essential for accurate interpretation and analysis.