Showing 1–2 of 2 results for author: Kanta, A
-
Smarter Password Guessing Techniques Leveraging Contextual Information and OSINT
Authors:
Aikaterini Kanta,
Iwen Coisel,
Mark Scanlon
Abstract:
In recent decades, criminals have increasingly used the web to research, assist and perpetrate criminal behaviour. One of the most important ways in which law enforcement can battle this growing trend is through accessing pertinent information about suspects in a timely manner. A significant hindrance to this is the difficulty of accessing any system a suspect uses that requires authentication via…
▽ More
In recent decades, criminals have increasingly used the web to research, assist and perpetrate criminal behaviour. One of the most important ways in which law enforcement can battle this growing trend is through accessing pertinent information about suspects in a timely manner. A significant hindrance to this is the difficulty of accessing any system a suspect uses that requires authentication via password. Password guessing techniques generally consider common user behaviour while generating their passwords, as well as the password policy in place. Such techniques can offer a modest success rate considering a large/average population. However, they tend to fail when focusing on a single target -- especially when the latter is an educated user taking precautions as a savvy criminal would be expected to do. Open Source Intelligence is being increasingly leveraged by Law Enforcement in order to gain useful information about a suspect, but very little is currently being done to integrate this knowledge in an automated way within password cracking. The purpose of this research is to delve into the techniques that enable the gathering of the necessary context about a suspect and find ways to leverage this information within password guessing techniques.
△ Less
Submitted 2 December, 2020;
originally announced December 2020.
-
Improving Borderline Adulthood Facial Age Estimation through Ensemble Learning
Authors:
Felix Anda,
David Lillis,
Aikaterini Kanta,
Brett A. Becker,
Elias Bou-Harb,
Nhien-An Le-Khac,
Mark Scanlon
Abstract:
Achieving high performance for facial age estimation with subjects in the borderline between adulthood and non-adulthood has always been a challenge. Several studies have used different approaches from the age of a baby to an elder adult and different datasets have been employed to measure the mean absolute error (MAE) ranging between 1.47 to 8 years. The weakness of the algorithms specifically in…
▽ More
Achieving high performance for facial age estimation with subjects in the borderline between adulthood and non-adulthood has always been a challenge. Several studies have used different approaches from the age of a baby to an elder adult and different datasets have been employed to measure the mean absolute error (MAE) ranging between 1.47 to 8 years. The weakness of the algorithms specifically in the borderline has been a motivation for this paper. In our approach, we have developed an ensemble technique that improves the accuracy of underage estimation in conjunction with our deep learning model (DS13K) that has been fine-tuned on the Deep Expectation (DEX) model. We have achieved an accuracy of 68% for the age group 16 to 17 years old, which is 4 times better than the DEX accuracy for such age range. We also present an evaluation of existing cloud-based and offline facial age prediction services, such as Amazon Rekognition, Microsoft Azure Cognitive Services, How-Old.net and DEX.
△ Less
Submitted 2 July, 2019;
originally announced July 2019.
