Image and video source identification in the era of mobile devices and social media

This Thesis introduces three novel image and video datasets, taking into account different types of multimedia and its alterations caused by the exchange through popular social media platforms. The first one is MOSES mobile application, proposed as an elegant option for providing an expandable, up-to-date video database. While the initial MOSES dataset contains SDR (Standard Dynamic Range) videos, the second dataset, named VISION, combines both SDR and HDR (High Dynamic Range) images and videos, thus providing the ability of comparison of different types of multimedia acquired by the same device. Due to their rising popularity, special attention is paid to HDR images. The third proposed dataset is one of the largest currently available HDR-based datasets and it enables SDR and HDR images comparison.