|eForensics Magazine 2020 03 Face Recognition PREVIEW.pdf|
Facial recognition systems are now more common than ever. Many countries are deploying their mass surveillance system in an effort to combat the spread of the coronavirus. Millions of people are and will be subject to unprecedented monitoring. The governments say it is being used to keep everyone safe but concerns are growing over the "big brother" use of data, giving police the power to create a digital authoritarian state.
Of course facial recognition systems can be very useful - starting from facial phone locks or tagging in social media and ending with identification of genetic disorders or cracking down on detecting illegal arrivals in the country at the airport. But the biggest strength of the recognition technology could also prove to be its greatest weakness, as there are a lot of issues concerning privacy. In this issue, dedicated to the topic of face recognition, you will learn about face recognition systems and technology, security, face detection and analysis. After reading this magazine, nothing regarding this topic will surprise you anymore.
The edition starts with the article “Face recognition - Introduction” by Dr. P. Mary Jeyanthi & Abhishek Maurya in which authors explain what facial recognition really is, and introduces face detection, posing and projecting faces, and how to encode faces. Then, we have the article entitled “Face Recognition: The Good, The Bad, and The Ugly”. Damien E.B. Mallet shows us perfectly that this technology isn’t flawless.
Further, you can read about commonly available recognition technology in the paper prepared by Matthew Kafami. And then, there is a really extensive and interesting article by Jim Hoerricks, “The subjectivity of consistency, an overview of manual facial photographic comparison in forensic science”.
What’s also covered? Face recognition applications, face morphing attacks, face recognition - forensic modality, some tips regarding conducting an OSINT investigation, memory analysis of WannaCry ransomware, and memory analysis using Volatility and Ghidra.
Check out our Table of Contents below for more information about each article (we included short leads for you).
We hope that you enjoy reading this issue! As always, huge thanks to all the authors, reviewers, to our amazing proofreaders, and of course you, our readers, for staying with us! :)
Have a nice read!
and the eForensics Magazine Editorial Team
TABLE OF CONTENTS
Face Recognition - Introduction
by Dr. P. Mary Jeyanthi & Abhishek Maurya
How excited are we when we receive a Facebook notification that it has auto-detected our face in a friend’s picture? Or when the Google gallery app lets us organize the images by the person who is in them. These are basic everyday examples of face recognition. Despite how you feel about face recognition, it has now become a common way to authenticate people after fingerprint and iris recognition. While face recognition might not be as reliable as the other types of security authentication modes, it is widely adopted due to its contactless process. Before we discuss any more on face recognition, we need to understand what it really is.
Face Recognition: The Good, The Bad and The Ugly
by Damien E.B. Mallet
The human brain is an amazing tool to recognise patterns. We use its ability effortlessly every day and most of the time without even thinking about it. It is no wonder that researchers and engineers want to replicate this ability into machines and help automate certain processes. Nowadays, average computers can process billions of calculations per second. Their processing power is essential to compete with what our brain does in the blink of an eye. However, processing power alone is not enough to replicate what nature created through millions of years of evolution. Novel algorithms and techniques had to be developed in order to match and surpass human capabilities to some extent.
Commonly Available Recognition Technology
by Matthew Kafami
There are two major publicly accessible recognition technologies that the average person might use: facial recognition software and advanced license plate recognition, or ALPR.
The subjectivity of consistency, an overview of manual facial photographic comparison in forensic science
by Jim Hoerricks, PhD AVFA
Facial photographic comparison for the purpose of identifying people has been present in the world’s courts since at least the famous Tichborne case, heard in the English Court of Common Pleas in 1871-72. One hundred fifty years later, courts are still attempting to accurately interpret the presented results of facial identification examinations, with results and conclusions often offered in subjective language. In this paper, the current advice from several standards producing bodies, court rulings, and the body of literature are used to frame a discussion of the process of comparing images of faces and presenting findings, providing practitioners and evaluators a framework with which to assess presented evidence. An annotated case survey is provided in the references section for further information.
Face recognition - Applications
by Subhabrata Debnath
Facial recognition has become increasingly popular over the years owing to its wide array of applications and the ease with which it can be deployed. Availability of high definition cameras, fast internet and high-performance GPUs have all contributed to its development.
Face Morphing Attacks: A Threat to Facial Recognition Systems?
by Rhonda Johnson
While touted as a secure tool of smart surveillance technology, there are limitations to facial recognition systems that malicious actors can exploit. Creating a computational model or template of a human face is challenging because the human face is always changing. Aging, surgery modification, illness, captured lighting, facial expressions and physical changes such as trauma and weight gain can possibly alter the dimensions of a face, the same dimensions that are used to create probes or templates for comparison to a database. A facial morphing attack takes advantage of such vulnerabilities to manipulate the digital composition of a face in efforts to avoid detection by a facial recognition system. Face morphing combines more than one face image to generate a single image that can be used to trick the facial recognition system into evading image detection from a facial recognition image database.
Face Recognition: The Forensic Modality
by Zeeshan A Khan
Forensic science is the major application of technology to law enforcement, and the strong forensic evidence evaluation leads to identify a suspect efficiently. The identification of a perpetrator from the available evidence requires an in-depth analysis of the crime scene. In this regard, biometric modalities such as face recognition, fingerprint recognition, voice recognition, and gait recognition can provide exceptional help. Whereas, face recognition, due to its agility and accuracy, is the most effective modality.
Before conducting OSINT: privacy, security and browser tips
“The key word behind OSINT concept is information, and most importantly, information that can be obtained for free. It doesn't matter if it is located inside newspapers, blogs, web pages, tweets, social media cards, images, podcasts, or videos as long as it is public, free and legal.” – Securitytrails
Large File Analysis Using Volatility (Part III)
by Paulo Pereira, PhD
This is Part III of a long article in which I decided to analyze a large memory image captured from a 64 bits Windows 10 system to discover any suspicious behavior using Volatility. Commercial version Volcano (Volexity Incident Response and Threat Intelligence tool, please see the reference) lead very well with a large memory file. But, in this article my aim is to use Volatility. Part III is also dedicated to the analysis of the dumped executable files of a specific process to make some analysis of suspicious behavior using radare2 and ghidra.
Memory Analysis of WannaCry Ransomware using Volatility
by Sumit Kumar
In this article, we will discuss analyzing the RAM dump of an infected computer using the Volatility tool. Volatility is one of the best open source memory forensics framework for Incident Response and Malware Analysis. It is based on Python and supports analysis for Linux, Windows, Mac OS and Android systems. It can analyze raw dumps, crash dumps, VMware dumps (.vmem), virtual box dumps, and many others.