A deep learning model for splicing image detection

Abstract

With the growth of digital technology, modifying images using various photo editing techniques has become quite simple. One of the techniques is the splicing image method, which crops parts of images and puts them into another image creating a new composite image. The image splicing detection system is soon regarded as an exciting topic for many researchers to solve the problems of forgery images on the Internet, especially in social networks. Applying deep learning techniques in the task of detecting splicing images is a promising approach to combat image forgery and manipulation. By utilizing the strength of convolutional neural networks and their ability to understand complicated patterns and features, deep learning models may be trained to effectively recognize the presence of splicing in digital images. ResNet-50 and VGG-16 are powerful architectures of convolutional neural networks, but they reveal many weaknesses when operating on low-end computers. The ultimate goal of this research is to develop a model for image splicing detection that works effectively on machines with limited memory. The study proposes the model, which is the improvement of VGG-16 applying residual network (ResNet). The proposed method comprises a series of steps to detect and classify authentic and fake images. These steps involve initial dataset separation into the train, validation, and test sets, enabling effective model training and evaluation. In the pre-processing phase, an analysis is conducted on the level of compression error present in the images, helping to identify potential artifacts and anomalies introduced during manipulation. Subsequently, the dataset is utilized to train the model, leveraging deep learning techniques to learn discriminative features and patterns indicative of image authenticity. Finally, the trained model is evaluated, assessing its performance in accurately classifying images as either authentic or fake. This comprehensive methodology provides a systematic framework for robust image authentication and forgery detection. As a result, after 20 epochs of training 9,319 images from the CASIA v2.0 dataset, the proposed model achieves a test accuracy of 92.5%, while the ResNet-50 achieves a precision of 85.6%. The outcome demonstrates the effectiveness of the suggested model for image splicing detection, particularly when operating on low-end machines.