Kenyan sign language recognition using the ensemble method.

Abstract

Communication is the process of interaction between two or more individuals. It can be done through verbal, written, gestures such as hand and head movement, facial expression, lip motion. Communication is fundamental to social inclusion, yet the Deaf community in Kenya continues to face significant barriers due to limited recognition and translation of Kenyan Sign Language(KSL). Unlike American or Indian Sign Language, KSL has unique linguistic and cultural structures that make generic recognition systems unsuitable. This study addresses these challenges by developing an ensemble machine-learning model for KSL recognition that combines the feature extraction power of Convolutional Neural Networks (CNN) with the classification robustness of k-Nearest Neighbors (KNN). To enhance diversity and reduce noise the model was preprocessed, augmented, and annotated with a curated dataset of 8,898 KSL images, which were obtained from Kaggle. CNN extracted high-level spatial features from the gesture images, and the KNN classifier used this information for similarity-based decision making. The ensemble strategy was used to combine the outputs of both models to enhance accuracy and reduce the ability to be misclassified. Precision, recall, F1-score, and confusion matrices were used to evaluate the performance on the test set. This ensemble model achieved validation accuracy of 70.32% outperforming standalone models and demonstrates improved recognition of KSL gestures and notations. The findings of this study highlight how ensemble learning can bridge communication gap between the Deaf and the hearing world in Kenya. The research not only gives a technological solution to real-time KSL recognition but also lays a foundation for future research on larger, more diverse datasets and dynamic gesture recognition. Finally, this work adds to the social inclusion of the Deaf community by facilitating them access communication tools that empower and provide them with equal access to education, healthcare, and everyday life.