TITLE:
Development of a System for the Recognition of Isolated Signs in Text of the Niger Sign Language (LSNi) Based on Neural Networks
AUTHORS:
Bachir Moussa Idi, Chaibou Kadri, Ibrahim Bouwey Alley, Yahaya Morou Ganda, Harouna Naroua
KEYWORDS:
Niger Sign Language, Automatic Translation, Recurrent Neural Networks, LSTM
JOURNAL NAME:
Journal of Computer and Communications,
Vol.14 No.9,
September
16,
2026
ABSTRACT: This article contributes to improving communication between the deaf community and the rest of the population through the use of digital technology and artificial intelligence techniques, notably Deep Learning, by developing a system capable of recognizing isolated signs in text of Niger Sign Language (NiSL) gestures (words) in real time using a webcam. For this purpose, a video dataset specific to Niger Sign Language was created. Landmarks of articulations were extracted using MediaPipe Holistic [1] and used to train a Long Short-Term Memory (LSTM) classification model to perform this recognition of isolated signs in text. The results obtained show a good performance of the model, with a recognition rate of 62.22% when only hand characteristics are used, which is comparable to results of previous works in the literature on more documented sign languages.