Web Page Harvesting for Automatized Large-scale Digital Images Anomaly Detection

Abstract

Currently, digital media content is increasingly being used by cybercriminals for nefarious purposes. Such objects can be used, e.g., to covertly transfer malicious code to the infected host or to exfiltrate sensitive information from the secured perimeter to the attacker’s server. In this paper, we present the design and deployment of a web page harvesting platform that allows performing various types of large-scale analyses, including metadata inspection, detection of hidden data, or evaluation of compliance with the graphical standard. The platform architecture has a distributed, flexible, and modular form, making it easily extendable and efficient. In this article, we also include initial experimental results of the analyzes carried out on the content of 1,000 of the most popular websites.