LLM-Assisted Behavioral Biometric Authentication Reaches 99.46% Accuracy
Original title: LABBA: LLM-Assisted Behavioral Biometric Authentication
LLM reasoning boosts phone-blowing biometrics to 99.46% accuracy on 50 participants.
The framework integrates an LLM-based reasoning module with a novel blow-acoustic biometric technique. Spectral and temporal features are extracted from signals and converted into structured prompts via prompt engineering. Authentication accuracy reaches 99.46% in rule-based mode and 99.38% in score-based mode while maintaining robustness under behavioral and contextual variability.
- Accuracy reaches 99.46% in rule-based mode.
- Outperforms non-LLM counterpart.
- Tested with 50 participants.
- Robust to dynamic real-world conditions.
From the original article
1 min read at arxiv.orgBehavioral biometrics have emerged as a promising complement to traditional physiological biometrics by leveraging dynamic human characteristics. These modalities offer advantages including continuous authentication, inherent liveness detection, and greater resistance to credential theft and other spoofing attacks.
By Howard Halim · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.