Mark Scanlon, Ph.D. (University College Dublin)
Dr. Katrin Franke
Zeno Geradts (Netherlands Forensic Institute)


Link to downloads:


This workshop covers varied aspects of data science and machine learning in its relationship to digital forensics and policing.  The content in this workshop ranges from beginning material to intermediate tutorials. 


Zeno Geradts of the Netherlands Forensic Institute kicks off the tutorial with a presentation of the challenges and opportunities related to forensic intelligence from the perspective of law enforcement. The following presentation by Katrin Franke of NTNU lays out the fundamentals of machine learning, wherein the basic concepts of features and models are explained, and the techniques of classification, clustering, and regression are introduced.  This first half of the workshop prepares and motivates the hands-on second half.


Our hands-on tutorials begin with a walk through of a basic data science example for classification.  This starts with preparing our dataset, which includes data visualization and data preprocessing.  Subsequently we use our prepared data to train and test our machine-based classifier. This example uses a common dataset for beginners, and allows for hands-on experience with Support Vector Machines (SVM) and Neural Networks for classification.


The second part of the hands-on workshop will give an overview of the state of the art for computer vision/deep learning based facial age estimation.  Attendees will be given sample code to get a real-time age and gender estimation system running in their machines.  The instructions for installing the prerequisite software are given on the workshop website. 

Brief Bio's


Dr. Mark Scanlon is an Assistant Professor in the UCD School of Computer Science and Founding Director of the UCD Forensics and Security Research Group. He is a Fulbright Scholar in Cybersecurity and Cybercrime Investigation. Both his MSc and PhD are in the field of Remote Digital Forensic Evidence Acquisition. His research interests include Remote Evidence Acquisition, Evidence Whitelisting & Data Deduplication, Cloud Forensics, File Synchronization Service Forensics, Network Forensics, and Digital Forensics Education. Dr. Scanlon is an active member of the digital forensics research community and is a keen editor, reviewer and conference organizer across a range of key journals and conferences in the field including Digital Investigation, the Journal of Digital Forensics, Security and Law, and DFRWS.


Dr. Mark Scanlon is an Assistant Professor in the UCD School of Computer Science, Principal Investigator for Cybersecurity at the \euro12m funded Centre for Applied Data Analytics Research and Machine Learning (CeADAR), Founding Director of the UCD Forensics and Security Research Group, and the Director of the School's Masters in Forensic Computing and Cybercrime Investigation. His research interests include Automated Evidence Processing, Evidence Whitelisting & Data Deduplication, Intelligent Systems for Cybersecurity, and Network Forensics and Analytics. Dr. Scanlon is an extremely active member of the digital forensics and cybersecurity research communities and is a keynote speaker, associate editor, reviewer, and conference organiser across a range of key journals and conferences in these fields.


Katrin Franke is a (full) professor in computer science within the information security environment at the Norwegian University of Science and Technology (NTNU) in Gjøvik. In 2007 she joined the Norwegian Information Security Lab (NISlab) with the mission to establish research and education in digital and computational forensics. In this context she was instrumental in setting up the partnership with the Norwegian police organisations as part of the Center for Cyber and information Security (CCIS). Dr. Franke is now leading the NTNU Digital Forensics group. Dr. Franke has 20+ years experiences in basic and applied research for financial services & law enforcement agencies (LEAs) working closely with banks and LEAs in Europe, North America and Asia.


Zeno Geradts is a senior forensic scientist at the Netherlands Forensic Institute of the Ministry of Justice at the Digital Evidence and Biometrics section in the area of forensic (video) image processing and biometrics within the team Forensic Big Data Analysis. From September 1st 2014, he is appointed as professor on forensic data science by special appointment at the University of Amsterdam. He is chairman of the Forensic IT working group of ENFSI  From February 2019 to February 2020 he is President of the American Academy of Forensic Sciences 


Kyle Porter is a PhD candidate and research fellow in the Testimon Digital Forensics Lab at the Norwegian University of Science and Technology. His primary research interest is to increase the efficiency and effectiveness of finding relevant textual data or evidence in digital investigations. The methods being investigated include developing string matching algorithms for literal text matches, or applications of machine learning for intelligent data exploration.


The instructions for installing the prerequisite software are given below:

(installation instructions copied below for reference)

Part 1: Basic Data Science Setup Instructions

Part 2: Facial Age Estimation Setup Instructions

Option 1 (recommended):

The quickest and easiest option for facial age estimation demonstration part of the workshop is to use the latest version of VirtualBox platform and the VirtualBox Extension Pack (both available from here: Then you can download a pre-prepared virtual machine (dfrws.ova ~1.4Gb) from either of the below links:

Google Drive:
UCD server:

dfrws.ova MD5 hash: aee02f23b2bed023b74b66885b9a787c
dfrws.ova SHA1 hash: 88d2c7fe5ec2d0639416f8c8b3d72687781fc61a

OS username and password: lubuntu/lubuntu

Option 2:

Manual installation instructions for linux (requires python 2.7.x):

0) sudo apt install git python-pip

1) git clone

2) cd dfrws_demo

3) sudo apt-get install build-essential cmake libgtk-3-dev libboost-all-dev

4) pip install -r requirements.txt

5) python (you may be prompted to grant python access to your webcam; the demo needs this access)

Note: It can take up to 15 minutes to download the weights file when launching for the first time.

This demo uses the WideResnet Architecture for images of size 64x64 trained on the IMDB-WIKI dataset. Reference: