Wearable Device-Based Gait Recognition Using Angle Embedded Gait Dynamic Images and a Convolutional Neural Network.

Area of Science:

• Biometrics
• Machine Learning
• Signal Processing

Background:

• Inertial sensors in wearable devices offer potential for gait recognition.
• Extracting discriminative features from noisy sensor data is challenging.

Purpose of the Study:

• To propose a novel image-based gait recognition approach using Convolutional Neural Networks (CNNs).
• To develop a method that does not require manual feature extraction for gait analysis.

Main Methods:

• Utilized Angle Embedded Gait Dynamic Image (AE-GDI) as a 2D representation of gait dynamics.
• AE-GDI is invariant to rotation and translation, encoded from inertial sensor data.
• Employed CNNs for gait authentication and labeling.

Main Results:

• The proposed approach achieved competitive recognition accuracy on two datasets (McGill and Osaka Universities).
• Demonstrated effectiveness in realistic conditions and with a large number of subjects.
• AE-GDI provides a robust representation for gait dynamics.

Conclusions:

• The novel image-based CNN approach offers an effective solution for gait recognition.
• AE-GDI is a promising representation for gait dynamics, simplifying feature extraction.
• The method is suitable for identifying individuals among large populations using wearable device data.