A Review on Image Forgery Detection Techniques Using Machine Learning

Image forgery has evolved into common problem in the digital age, due to the extensive uses of digital image manipulation tools. In a variety of industries, including forensics, journalism, and arts, image fraud can have detrimental effects.<br><br>Thus, it is crucial to provide trustworthy techniques for identifying image forgery. Using machine learning techniques to automatically spot indications of image modification is one promising strategy. We give a summary of current developments in machine learning-based image forgery detection in this review paper. We talk about many methods of forging images, including copy-move, splicing, and retouching. We also give an overview of common machine learning techniques used in picture forgery detection, including SVM, CNN and Random Forests. The performance of various features extraction techniques to capture the distinctive aspects of various types of image forgeries is then discussed, including the Scale-Invariant Feature Transform and convolutional neural network-based features. Several datasets that have been utilized to train and evaluate machine learning models for image forgery detection are also reviewed. Finally, we evaluate the shortcomings of current approaches and specify potential future research avenues. We stress the importance of creating reliable methods that can identify cutting-edge types of image forgery, such as deepfakes, as well as the necessity of creating real-time, practical solutions.<br> This review paper intends to be a helpful resource for scholars and practitioners working in this field by giving a thorough overview of recent developments in picture forgery detection using machine learning.