MeshID: Few-Shot Finger Gesture Based User Identification Using Orthogonal Signal Interference.

Area of Science:

• Human-Computer Interaction
• Signal Processing
• Biometrics

Background:

• Radio frequency (RF) technology offers device-free sensing for human-computer interaction.
• Challenges in RF-based identification include low signal resolution and user variability.
• Existing methods struggle with subtle finger gestures and complex environments.

Purpose of the Study:

• To propose MeshID, a novel RF-based scheme for accurate user identification using finger gestures.
• To enhance RF signal sensitivity for extracting subtle individual biometrics.
• To develop a robust and efficient few-shot model retraining framework.

Main Methods:

• Utilizing RF signal interference to extract velocity distribution profiling (VDP) features.
• Implementing a few-shot model retraining framework with a first component reverse module.
• Conducting comprehensive real-world experiments in diverse indoor environments.

Main Results:

• MeshID achieves an average user identification accuracy of 95.17% across three indoor settings.
• The system demonstrates high model robustness and performance in complex environments.
• MeshID outperforms state-of-the-art methods in identification accuracy and cost-efficiency.

Conclusions:

• MeshID offers a highly accurate and cost-effective solution for RF-based user identification.
• The velocity distribution profiling (VDP) feature extraction effectively captures subtle user biometrics.
• The proposed few-shot learning framework enhances model adaptability and performance.