Treffer: Transforming Image Super-Resolution through Redefined Generative Adversarial Network with Edge-Aware Loss Function (RGAN-EALF).

Image super-resolution is the technology dealing with the generation of high-quality images from lowresolution ones, an essential function in a number of applications. Most of the classic approaches cannot preserve the intricate details and sharp edges. Recent research developments in artificial intelligence raise some encouraging options. In this paper, the authors propose a new method to enhance the image quality and preserve the details using an innovative Redefined Generative Adversarial Network with an Edge-Aware Loss Function (RGAN-EALF) algorithm. This approach aims to upscale images while maintaining sharpness and clear edges. The model is trained and evaluated on a diverse set of high-resolution images, ensuring broad coverage of different textures and structures. Images are standardized and resized to the same resolution to maintain uniformity throughout the dataset. Histogram equalization and other techniques are employed to amplify contrast and improve the salience of features. Key features are extracted by CNN to identify and preserve important information and textures in the images. The RGAN consists of a generator and a discriminator, where the generator aims at generating high-resolution images. EALF is used to enhance image edges and preserve complex details by focusing on edge information during training. The proposed model is implemented using Python software. The performance of the proposed model is evaluated and compared to state-of-the-art techniques using traditional image quality metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index measure (SSIM), precision (98.86%), recall (97.70%), and f1-score (97.85%). Results reveal that the RGAN-EALF model provides higher quality images and preserves more image details. The model is applied to the WXCD dataset. It consists of multi-sensor high-resolution images designed to identify building changes. This dataset was captured by Unmanned Aerial Vehicle and SuperView-1 satellite images and has substantial variations between UAVs and satellite-captured images. The dataset processing has been done using ArcGIS software, manually annotating the data and converting it to the TIFF format. [ABSTRACT FROM AUTHOR]

Copyright of International Journal of Intelligent Engineering & Systems is the property of Intelligent Networks & Systems Society and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)