Related Experiment Video
Updated: Dec 24, 2025
08:15
Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision
Published on: March 28, 2025
1.1K
Dynamic Hand Gesture Recognition Based on a Leap Motion Controller and Two-Layer Bidirectional Recurrent Neural
Linchu Yang1, Jian Chen1, Weihang Zhu2
1Department of Mechanical Engineering; Jiangsu University of Science and Technology, Zhenjiang 212003, China.
Sensors (Basel, Switzerland)
|April 12, 2020
Summary
A novel two-layer Bidirectional Recurrent Neural Network enhances dynamic hand gesture recognition using Leap Motion Controller (LMC) data. This system achieves high accuracy on American Sign Language and Handicraft-Gesture datasets, improving human-computer interaction.
Area of Science:
- Computer Science
- Artificial Intelligence
- Human-Computer Interaction
Background:
- Dynamic hand gesture recognition is crucial for intuitive human-computer interaction.
- Existing methods require improvement in accuracy and efficiency.
- Leap Motion Controller (LMC) offers a viable platform for gesture data capture.
Purpose of the Study:
- To propose an accurate and efficient dynamic hand gesture recognition system.
- To leverage a two-layer Bidirectional Recurrent Neural Network (BiRNN) for improved recognition.
- To validate the system's performance on diverse gesture datasets.
Main Methods:
- Development of a two-layer BiRNN model for dynamic hand gesture recognition.
- Utilizing feature vectors extracted from Leap Motion Controller (LMC) data.
- Testing the system on American Sign Language (ASL) and Handicraft-Gesture datasets.
Main Results:
- Achieved 100% training accuracy and over 95% testing accuracy on ASL datasets (360 and 480 samples).
- Attained 100% training accuracy and 96.7% testing accuracy on the Handicraft-Gesture dataset.
- Cross-validation results (5-fold, 10-fold, Leave-One-Out) demonstrated robust performance across datasets.
Conclusions:
- The proposed two-layer BiRNN system effectively recognizes dynamic hand gestures from LMC data.
- The system demonstrates high accuracy and efficiency, comparable or superior to existing methods.
- This approach offers a significant advancement in human-computer interaction through gesture recognition.
