Keyboard acoustic side channel attacks exploit audio leakage from typing to deduce typed words with a certain level of accuracy. Researchers have been improving the accuracy of these attacks through various techniques of feature extraction and classification. They also apply machine learning methods and deep learning techniques to enhance the accuracy of their results. At the same time, defense mechanisms against these attacks have not kept up with the increasing precision of the attacks. In this study, we introduce a practical defense strategy against keyboard acoustic attacks during password and text typing. We evaluate its effectiveness against multiple attack vectors. Our defense strategy involves generating unique background sounds using GANs to mask sensitive audio leaks from the keyboard, thwarting side channel attacks from extracting usable information about typed content. The background sounds are produced by the device used for text input. We assess the usability of our approach for short and prolonged usage durations, demonstrating that the addition of background sounds does not impede users’ ability to enter passwords or perform computer tasks effectively.
Online social networks serve as platforms facilitating communication, collaboration, and information exchange among hackers. Consequently, the detection and monitoring of hacker communities within these networks have emerged as significant research areas in the field of cybersecurity. Numerous methods have been employed to detect these communities, with some primarily utilizing network structural information and others incorporating user-generated content topics. However, most existing methods rely on probabilistic models such as TF-IDF or LDA to analyze textual similarity, resulting in limited accuracy. In this paper, we introduce a Siamese Attention-augmented Recurrent Convolutional Neural Network aimed at analyzing text similarity among users and evaluating the significance of different users by comparing their text similarity with seed users. Subsequently, a weighted network is constructed, and community detection is performed using a modularity maximization algorithm. To validate the proposed framework, we collected Twitter friendship data and user-generated content from a pre-existing regional hacker group. A comprehensive series of experiments were conducted on this dataset. The experimental results provide strong evidence of the effectiveness of our proposed framework, showcasing its remarkable performance across various indicators such as community quality and topic relevance. Consequently, our framework exhibits promising prospects for practical applications in the field of hacker community detection.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.