Implementation of Convolutional Neural Network for Emergency Sound Detection for Hearing-Impaired Individuals on Android

Muhammad Akram Fais (State University of Medan)
Insan Taufik (State University of Medan)
Mansur AS (State University of Medan)
Debi Yandra Niska (State University of Medan)
Hanna Dewi Marina Hutabarat (State University of Medan)

Journal of Artificial Intelligence and Engineering Applications (JAIEA)
Vol. 5 No. 3 (2026): June 2026

Hearing impairment is a condition characterized by partial or total loss of hearing ability, which may occur congenitally or be caused by factors such as injury, disease, or prolonged exposure to excessive noise. This study aims to develop an Android-based emergency sound detection system using the Convolutional Neural Network (CNN) method. The research workflow includes problem identification, data collection, data preprocessing, CNN model training, model evaluation, Android application development, and system testing. Experimental results show that the best-performing model achieved an overall accuracy of 93%. The trained model was then implemented into an Android application to enable real-time sound classification and to provide visual notifications when emergency sounds are detected. The evaluation results indicate that the CNN model is capable of accurately classifying emergency sounds and operates effectively on Android devices.