Related Experiment Video
Updated: May 21, 2025
08:15
Capturing Dynamic Finger Gesturing with High-resolution Surface Electromyography and Computer Vision
Published on: March 28, 2025
333
Innovative hand pose based sign language recognition using hybrid metaheuristic optimization algorithms with deep
Bayan Alabduallah1, Reham Al Dayil2, Abdulwhab Alkharashi3
1Department of Information Systems, College of Computer and Information Sciences, Princess Nourah Bint Abdulrahman University, P.O. Box 84428, 11671, Riyadh, Saudi Arabia. Bialabdullah@pnu.edu.sa.
Scientific Reports
|March 19, 2025
Summary
This study introduces an innovative sign language recognition technique using hand pose estimation and deep learning. The method achieves 99.57% accuracy, significantly improving communication for hearing-impaired individuals.
Area of Science:
- Computer Science
- Artificial Intelligence
- Human-Computer Interaction
Background:
- Sign language (SL) is a vital visual-physical communication method for the deaf community.
- Sign language recognition (SLR) is crucial for bridging communication gaps between deaf and hearing individuals.
- Deep learning (DL) approaches have shown promise in advancing SLR capabilities.
Purpose of the Study:
- To present an Innovative Sign Language Recognition using Hand Pose with Hybrid Metaheuristic Optimization Algorithms in Deep Learning (ISLRHP-HMOADL) technique.
- To enhance the efficiency and accuracy of sign interpretation for hearing-impaired persons through improved hand pose recognition.
Main Methods:
- Image pre-processing using a wiener filter (WF) for noise reduction and quality enhancement.
- Feature extraction via a fusion of ResNeXt101, VGG19, and vision transformer (ViT) models.
- Hand pose recognition using a bidirectional gated recurrent unit (BiGRU) classifier.
- Parameter optimization using a hybrid crow search-improved grey wolf optimization (CS-IGWO) algorithm.
Main Results:
- The ISLRHP-HMOADL model demonstrated superior performance on the ASL alphabet dataset.
- Achieved a high classification accuracy of 99.57%, outperforming existing methods.
- The hybrid optimization algorithm effectively tuned model parameters for enhanced robustness and accuracy.
Conclusions:
- The proposed ISLRHP-HMOADL technique significantly advances sign language recognition capabilities.
- The fusion of advanced DL models and metaheuristic optimization offers a robust solution for hearing-impaired communication.
- This approach holds potential for developing more accessible and effective assistive technologies.
