Video-Based Cryptanalysis: Extracting Keys from Power LEDs
Recovers cryptographic keys by analyzing video footage of device power LEDs, using camera rolling shutters to measure CPU power consumption fluctuations.
Summary
Optical Side-Channels via Power LEDs
Cryptographic operations induce subtle power consumption fluctuations within processors, which instantaneously alter the brightness and color spectrum emitted by device power indicator LEDs.
High-Frequency Sampling via Camera Rolling Shutters
Ben Nassi and Ofek Vayner demonstrate that commercial video cameras with rolling shutters can sample LED luminosity across individual sensor scan lines, boosting effective optical sampling rates by 1000x.
Extracting 256-Bit ECDSA and SIKE Keys at Distance
Experimental validations demonstrate extracting a 256-bit ECDSA private key from a smartcard reader from 16 meters away via a security camera, and recovering post-quantum SIKE keys from smartphone peripherals.
Countermeasures and Physical Mitigations
Mitigating optical cryptanalysis requires hardware filtering capacitors, constant-current LED drivers, physical light diffusers, and constant-time cryptographic libraries.
Related talks
Main Stage·1 hr 17 min
Reza Javadi · Simon Levy · Aaron Draffe
Full reverse engineering of HID's iCLASS SE chain of trust and extracting master cryptographic keys from CC EAL 5+ secure elements.
Main Stage·40 min
Elie Bursztein
Demonstrates how deep learning models outperform classical DPA attacks for power-analysis-based key extraction from AES implementations.
Main Stage·43 min
Yaniv Balmas · Lior Oppenheim
Yaniv Balmas and Lior Oppenheim demonstrate reverse engineering KVM switch firmware, subverting the 8051 microcontroller to install an undetectable, persistent hardware keylogger.