International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 08|Aug 2023
p-ISSN: 2395-0072
www.irjet.net
A Novel Approach for Enhancing Image Copy Detection with Robust Machine Learning Techniques 1
Prof.Pushpalata Patil , Limbale Laxmi Mallinath(Surekha) 2 1
Professor, Dept. of Computer Science and Engineering, Sharnbasva University, Kalaburagi ,Karnataka ,India Student, Dept. of Artificial Intelligence and Data Science, Sharnbasva University, Kalaburagi ,Karnataka ,India ----------------------------------------------------------------------------------***---------------------------------------------------------------------------------images are circulated without verification. The rapid Abstract advancement of image editing software and the ubiquity of high-quality cameras exacerbate these concerns, making it Digital Image Forgery involves manipulating images to increasingly difficult to discern authentic images from obscure or alter important content, often making it manipulated ones. To address these challenges, effective and difficult to identify tampered areas. Detecting and robust methods for detecting and localizing digital image preventing such manipulations are essential to maintain forgery are imperative. This project delves into the realm of the credibility of images, especially in an era where image forensics, focusing on the development and evaluation advanced photography tools and editing software enable of techniques that can accurately identify manipulated easy exploitation. This paper focuses on surveying diverse regions within images. By leveraging machine learning forgery types and methods for detecting digital image algorithms and optimization processes, we aim to contribute manipulation. This study employs the Particle Bee Firefly to the growing body of research aimed at safeguarding the Optimization Algorithm (PBFOA) to enhance feature credibility of digital imagery. The following sections of this extraction through component analysis. Additionally, Fuzzy paper provide a comprehensive overview of the project's C-Means (FCM) is applied for image segmentation. PBFOA objectives, methodology, experimental setup, and results. By aids in selecting valuable features by evaluating fitness exploring various forgery detection techniques and functions. Furthermore, the project adapts the U2-Net evaluating their performance, we seek to advance the field of model for image forensics and conducts experiments digital image forensics and provide practical solutions for comparing its effectiveness with ManTra-Net, another mitigating the adverse effects of image manipulation. forgery detection model. The experimental outcomes Through our research, we aspire to enhance the reliability of underscore U2-Net's versatility, showcasing its proficiency visual information in the digital landscape and promote the not only in identifying image forgery but also in accurately ethical use of digital images. Throughout this project, we pinpointing manipulated regions within images. explore the capabilities of the Particle Bee Firefly Intriguingly, U2-Net surpasses ManTra-Net in localized Optimization Algorithm (PBFOA) in conjunction with feature forgery detection in certain scenarios. This research sheds extraction and image segmentation techniques. By leveraging light on the complexities of digital image forgery detection PBFOA for feature selection and extraction, coupled with U2and highlights the potential of U2-Net as a potent tool for Net's potential in image manipulation detection and maintaining the authenticity and credibility of digital localization, we aim to contribute to the expanding toolkit of images. forgery detection methods. Our experimental setup Keywords: Digital Image Forgery, Image encompasses the implementation of PBFOA and U2-Net in Manipulation, Forgery Detection, Image Authenticity, conjunction with comparative analyses against established Image Integrity, U2-Net. models like ManTra-Net. Through this research, we aim to shed light on the capabilities and limitations of these 1. INTRODUCTION approaches, offering insights into their potential real-world applications. In today's digital age, images serve as powerful tools for communication, information dissemination, and artistic 2. Related Works expression. However, this accessibility and ease of sharing also bring about challenges, particularly in the Article[1]"Emerging Trends in Deep Learning for Image realm of digital image forgery. Digital image forgery Forgery Detection: A Comprehensive Survey" by Jessica involves deliberate alterations to images, which can Miller in 2021 deceive viewers and compromise the integrity of visual Jessica Miller's survey delves into the dynamic landscape of content. Such manipulations range from subtle deep learning techniques applied to image forgery detection. retouching to more sophisticated operations like object The review explores advancements in convolutional neural removal, addition, or size modification. The consequences networks (CNNs), generative adversarial networks (GANs), of undetected image forgery can be far-reaching. and their variants for uncovering manipulated content. By Misinformation, deceit, and copyright infringement are assessing the trade-offs between accuracy and efficiency, the just a few of the issues that arise when manipulated 2
© 2023, IRJET
Impact Factor value: 8.266
ISO 9001:2008 Certified Journal
|
Page 462