Hybrid Deep Learning Model for Copy Move Image Forgery Detection
Deipthan Prabakar, R. Ganesan, D. Leela Rani, Neti Praveen, N. Kalyani, Shreesha Kalkoor Mudradi
Abstract
Digital image processing software has progressed to the point where it is trivial to create forgeries by employing various manipulative tactics on the actual (real) images. Professions including law, health, and education are vulnerable to modified images. Hence, it is essential to be able to determine whether or not an image is fake. One form of image fraud is called “copy-move,” in which a key object or objects are hidden from view by copying and pasting them into the same image. Many Copy-Move Forgery Detection (CMFD) models are rendered useless when the copied section contains noise or is resized before being pasted. Our goal of this research was to develop a highly effective and efficient detection method for this type of image forgery by utilizing a hybrid Deep Learning (DL) architecture. Firstly, the sample images are taken from MICCF2000. Second, alter the size of the images, and a filtering technique is used to eliminate any noise that may have been present in the original image. Finally, we combine Convolutional Neural Networks (CNN) and Support Vector Machine (SVM) to build a hybrid DL model. Measures such as accuracy, True Positive (TPR) and Negative Rate (TNR), False Positive Rate (FPR) and Negative Rate (FNR), Precision, and F1-score are used to verify the developed hybrid DL model. Using the proposed hybrid DL model, we were able to detect fake images with an accuracy of 95%.