Development of a Low-Cost Security System Based on Voice Recognition Using Artificial Intelligence

Condori W.J.L.; Macedo E.J.M.; Huamani A.L.P.; Talavera J.S.; Galdos J. · SSRG International Journal of Electrical and Electronics Engineering (2024) · DOI: 10.14445/23488379/ijeee-v11i6p135

Resumen

Voice recognition has been widely used in various applications, especially in the field of security. In this paper, we propose the development of a low-cost security system based on voice recognition using artificial intelligence. The system utilizes a Raspberry Pi 4B as a microcontroller and Python as a programming language. The system works with a pre-recorded database of voices from 20 people, and the new user’s voice is matched against the pre-recorded voices using Gaussian Mixture Model (GMM). We extracted Mel-Frequency Cepstral Coefficients (MFCC) from the recorded voices, which were used to train the GMM. The system achieved an accuracy rate of 95.42%, with an equal error rate of 4.57%. The proposed system is low-cost and easy to use, making it accessible to a wider audience. However, it has some limitations, such as only being able to work with a pre-recorded database of voices.

Instituciones: Universidad Nacional de San Agustín de Arequipa