An offline parallel architecture for forensic multimedia classification

Abstract Nowadays, the volume of multimedia heterogeneous evidence presented for digital forensic analysis has significantly increased, thus requiring application big data technologies, cloud-based forensics services, as well Machine Learning (ML) techniques. In domain, ML algorithms have been applied cybercrime investigation such child abuse investigations, malware classification, and image forensics. This paper addresses this issues deals with images videos. particular, work aims at proposing a classification tool parallel software architecture fast inspection, which is easy to use (to be used by officers during search), requires limited hardware resources it built on an open-source limit its costs. Moreover, must able quickly inspect multiple devices time. When positives are found in device, device will seized deeper later lab. It not otherwise, reducing inconvenience suspect time required next phase. As case study, we focus identification pornography images. Experimental results show that proposed capable guaranteeing high recall, process performances real scenarios.