Multiple Image Tampering Detection using Deep Learning Algorithm
Naga Sri Saiee Gadiparthi, Jyothik Sandeep Kadha, Venkata Dinesh Reddy Palagiri, Gilsey Chadalavada, Gopal Krishna Kumba, Cristin Rajan · 2023
Image manipulation has become a widespread problem with digital photos in recent years. The detection of copy-move picture forgeries, image splicing, and recoloring are the three basic types of image forgeries that are mentioned. The dataset for detection of copy-move images is MICC-220 which involves 220 images with varying lighting conditions and camerasettings. In this paper, Scale-invariant feature transform, DBSCAN algorithm for copy-move image detection, and a deep architecture of a convolutional neural network are some of the approaches and models used to identify recolored images. The dataset used for detection of imagesplicing is CASIA V2 dataset, which contains 4795 photos, to classify altered images and detect different types of image tampering. In addition, image Error Level Analysis, an image compression techniquesareinvolved, along with the convolutional neural network to identify the altered images.