Detection of False Replication of Digital Copy Using Hog and SVM Classification

The widely used technique of altering digital photographs nowadays is the copy-move technique. Copy-move forgery recognition is known to be successful when using key point-based detection (CMFD). This approach reduces the high temporal complexity in matching brought on by the high - dimensional of SIFT by clustering the main points according to scale and color, breaking them up into smaller clusters, and then matching each independently. The approach is superior to current state-of-the-art approaches in comparing time complexity and detection reliability by utilizing picture splicing and image sampling. The percentage of false positives was minimal, and Image Resampling Detection performed well across all datasets and a wide range of pictures. The exposure of image-based cybercrimes requires the discovery of such counterfeit pictures. In this digital world, research on forging photos and spotting them is promising. Considering the many picture kinds that users frequently deal with, the project, therefore, offers a perfect platform from which they may choose the best image fraud detection technique. The objective of this work is to identify and draw attention to the lawlessness committed with current digital photos. By using the SVM technique we get an accuracy of 97.5%. Although metadata material might be changed, rendering it untrustworthy on its own, it is employed here as a supportive parameter for the choice of the error level analysis. By spotting counterfeit pictures, people may avoid using them to trick or hurt others.