Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
GIDSL: Indian-Gujarati Isolated Dynamic Sign Language Recognition Using Deep Learning
Ist Teil von
SN computer science, 2024-06, Vol.5 (5), p.527, Article 527
Ort / Verlag
Singapore: Springer Nature Singapore
Erscheinungsjahr
2024
Link zum Volltext
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
In today's connected world, effective communication is crucial. However, due to limited familiarity with sign language, the deaf and mute community often struggles with communication barriers. The state of Gujarat with over 60 million population, has a sizable population affected by deafness and muteness as revealed by the Indian census. To solve this challenge, we conducted a pioneering study of dynamic Gujarati Sign Language (GSL) recognition. Our study introduces a three-step approach to comprehensive sign language recognition. First, a robust dataset that has been accurately created using various signs important in everyday life. We then use the Mediapipe algorithm to extract key features and significantly reduce space complexity. We then build a novel 6 layer deep learning model consisting of Long Short-Term Memory (LSTM) and dense layers. In the last phase, we developed a real-time detection system and subjected it to rigorous testing on a wide data set. The performance evaluation included metrics such as accuracy, recall, F1 score and precision. The experimental results demonstrate the high accuracy of our proposed algorithm compared to state of the art Indian dynamic sign language recognition methods. This research not only fills a significant research gap, but also inspires hope to improve communication accessibility within the deaf community and chart a transformative trajectory toward inclusion.