Video Sequence Segmentation Based on K-Means in Air-Gap Data Transmission for a Cluttered Environment

Area of Science:

• Cybersecurity
• Information Security
• Data Transmission

Background:

• Air-gapped systems enhance information security but remain vulnerable to unconventional data exfiltration.
• Computer screen brightness modulation presents a covert channel for data leakage, detectable remotely.
• Existing methods for data retrieval from screen modulation have limitations, particularly in noisy environments.

Purpose of the Study:

• To develop a new, robust method for automatic video sequence segmentation to retrieve data transmitted via screen brightness modulation.
• To overcome the drawbacks of previous pixel-based growth methods.
• To enable data recovery from modulated reflections and afterglow, even without direct screen visibility.

Main Methods:

• Utilized a high-speed camera (380 fps) for image acquisition.
• Employed amplitude spectrum characteristics of individual pixels and k-means clustering for pixel grouping.
• Applied averaging of values within clustered areas to recover the transmitted signal.

Main Results:

• Successfully recovered 2-PAM (pulse-amplitude modulation) signals despite 1000x greater interference levels.
• Demonstrated effectiveness even with modulated reflections and monitor afterglow.
• The method does not necessitate high-quality lenses for image acquisition.

Conclusions:

• The proposed method offers a significant advancement in retrieving data from covert channels based on screen brightness modulation.
• It provides a more resilient and accessible approach to data recovery from air-gapped systems.
• This technique enhances cybersecurity by offering new ways to detect and analyze potential data exfiltration attempts.