Related Experiment Video
Updated: Aug 23, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.7K
Sign language recognition by means of common spatial patterns: An analysis
Itsaso Rodríguez-Moreno1, José María Martínez-Otzeta1, Izaro Goienetxea1
1Department of Computer Science and Artificial Intelligence, University of the Basque Country (UPV/EHU), Donostia-San Sebastián, Spain.
Plos One
|October 31, 2022
Summary
This study introduces an Argentinian Sign Language (LSA) recognition system using hand landmarks and Common Spatial Patterns (CSP) for improved communication. The system achieved high accuracy, aiding deaf and hard-of-hearing individuals.
Area of Science:
- Computer Science
- Artificial Intelligence
- Linguistics
Background:
- Millions face communication barriers due to hearing loss, highlighting the need for accessible solutions.
- Existing communication methods for deaf and hard-of-hearing individuals often rely on interpreters, limiting spontaneous interaction.
- Developing automated sign language recognition systems can bridge this gap.
Purpose of the Study:
- To develop and evaluate an Argentinian Sign Language (LSA) recognition system.
- To address the communication challenges faced by LSA signers and non-signers.
- To explore the efficacy of hand landmark analysis and machine learning for LSA recognition.
Main Methods:
- Utilized the LSA64 dataset, extracting hand landmarks from video data.
- Applied the Common Spatial Patterns (CSP) algorithm for dimensionality reduction of landmark signals.
- Employed classifiers including Random Forest (RF), K-Nearest Neighbors (KNN), and Multilayer Perceptron (MLP).
Main Results:
- Achieved high accuracy rates between 0.90 and 0.95 for recognizing 42 distinct LSA signs.
- Demonstrated the effectiveness of CSP in feature extraction for sign language recognition.
- Validated the performance of RF, KNN, and MLP classifiers on the processed data.
Conclusions:
- The developed LSA recognition system shows significant promise in facilitating communication.
- Hand landmark analysis combined with CSP and machine learning offers a viable approach for sign language recognition.
- This technology has the potential to enhance accessibility for the deaf and hard-of-hearing community.
