Detection of Spliced Images in Social Media Application
Munera A. Jabaar, Saad Najim AlSaad · 2021
Social media in the current technological era plays a major role in people's daily lives. Every day hundreds of thousands of images are circulated on social media applications such as WhatsApp, Instagram, Twitter, Facebook, and Snapchat. Photo is one of the most popular types of media that is shared among users on social media. It has become easy for small groups and even for individuals to edit and manipulate these images on a large scale in a very short time in such a way threatening the credibility of these images. In this paper, a detection system is implemented for verifying and classifying the content of social media images. The system adopted Deep Learning based on a convolutional neural network (CNN) to detect spliced images on WhatsApp. The images in dataset CASIA v2 (transformed to be appropriate for WhatsApp) are used for training and testing. The results point to an accuracy of 99.19% of training and 87.438% of testing.