Abstract:Abstract Digitization of every daily procedure requires trustworthy verification schemes. People tend to overlook the security of the passwords they use, i.e. they use the same password on different occasions, they neglect to change them periodically or they often forget them. This raises a major security issue, especially for elderly people who are not familiar with modern technology and its risks and challenges. To overcome these drawbacks, biometric factors were utilized, and nowadays, they have been widely adopted due to their convenience of use and ease of hardware installation. Many biometric-based authentication schemes were proposed, but despite the advantages that they offer, recent research has shown that biometrics by itself cannot be considered as an inviolable technique. Recently, we have proposed StegoPass, a novel method that obtains the 68 facial points of a user and utilizes them as a stego message to an image. This produced stego image was the password. Although the experiments conducted showed maximum security, it would be challenging to enhance the robustness of the proposed model for even more attacks. This paper examines the utilization of multimodal biometrics as the secret message embedded in the image. More specifically, besides the extraction of the facial points, we extract the unique minutiae moments and combine them in a feature vector. This feature vector is then embedded in the image. Two different datasets were used, and the security of the method was tested against state-of-the-art deep learning models, i.e. generative adversarial networks, to test whether the image could be digitally synthesized and fool the proposed verification scheme. The results proved that the new enhanced version of StegoPass offers an extremely secure method as its predecessor.

Haptic Passwords.

G-PPG: A Gesture-related PPG-based Two-Factor Authentication for Wearable Devices.

KinWrite: Handwriting-Based Authentication Using Kinect.

A New Two-Factor Authentication Method Leveraging Vibration Feedback from Hand Gestures

Neuromuscular Password-Based User Authentication

Enhancing IoT Security via Cancelable HD-sEMG-Based Biometric Authentication Password, Encoded by Gesture

Brain Password: A Secure and Truly Cancelable Brain Biometrics for Smart Headwear.

TouchPass: Towards Behavior-irrelevant on-touch User Authentication on Smartphones Leveraging Vibrations

Utilizing binary code to improve usability of pressure-based authentication

Combating Password Vulnerability with Keystroke Dynamics Featured by WiFi Sensing

A Dynamic Encryption Interface Based on a Multimodal Visuotactile Sensor

Adaptive Human-Machine Interactive Behavior Analysis With Wrist-Worn Devices for Password Inference

Two-Factor Authentication for Keyless Entry System via Finger-Induced Vibrations

Secure User Authentication Leveraging Keystroke Dynamics via Wi-Fi Sensing

H2B: Heartbeat-based Secret Key Generation Using Piezo Vibration Sensors

Exploiting multimodal biometrics for enhancing password security

Measuring Neuromuscular Electrophysiological Activities to Decode HD-sEMG Biometrics for Cross-Application Discrepant Personal Identification With Unknown Identities

BioTouchPass: Handwritten Passwords for Touchscreen Biometrics

A New Approach for User Authentication Based on Biometrics

Towards Nonintrusive and Secure Mobile Two-Factor Authentication on Wearables

Enabling Finger-touch-based Mobile User Authentication via Physical Vibrations on IoT Devices