EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Sign Language Recognition Using Residual Network Architectures for Alphabet and Diagraph Classification

Martins E. Irhebhude, Adeola O. Kolawole, Wali M. Zubair

Journal of Computing and Social Informatics · 19.12.2024 · 4 (1) · 11-25

DOI: 10.33736/jcsi.7986.2025 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Computing and Social Informatics, online erschienen am 19.12.2024. Verfasst von Martins E. Irhebhude und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Communication is crucial in human life, enabling the exchange of information through various methods beyond spoken language. Sign language translation is crucial for bridging communication gaps between hearing-impaired and hearing individuals, promoting effective interaction and understanding. This study presents a comprehensive model for identifying alphabet and digraph signs using feature extraction techniques from ResNet architectures, specifically ResNet18, ResNet50, and ResNet101. The system was designed to integrate both hand gestures and facial expressions, enhancing the accuracy of sign language recognition. Classification of sign language images into alphabet and diagraph categories was assessed using Support Vector Machine (SVM). The resulting classification accuracies were 61.7% for ResNet18, 64.5% for ResNet50, and 66.5% for ResNet101. The research results emphasize how deeper ResNet models are effective in improving recognition accuracy. This proposed model has significant implications for educational applications as it addresses attention-related challenges and aims to enhance student engagement in learning processes, thereby contributing to developing more inclusive educational environments.

Zitieren

APA

Irhebhude, M. E., Kolawole, A. O., & Zubair, W. M. (2024). Sign Language Recognition Using Residual Network Architectures for Alphabet and Diagraph Classification. Journal of Computing and Social Informatics, 4(1), 11-25. https://doi.org/10.33736/jcsi.7986.2025

Vancouver

Irhebhude ME, Kolawole AO, Zubair WM. Sign Language Recognition Using Residual Network Architectures for Alphabet and Diagraph Classification. Journal of Computing and Social Informatics. 2024;4(1):11-25. doi:10.33736/jcsi.7986.2025

BibTeX
@article{Irhebhude2024,
  author = {Irhebhude, Martins E. and Kolawole, Adeola O. and Zubair, Wali M.},
  title = {Sign Language Recognition Using Residual Network Architectures for Alphabet and Diagraph Classification},
  journal = {Journal of Computing and Social Informatics},
  year = {2024},
  volume = {4},
  number = {1},
  pages = {11--25},
  doi = {10.33736/jcsi.7986.2025}
}
Kostenlos starten →