Efficiency of the DenseNet convolutional neural network model for detecting drowsiness in drivers
Article Sidebar
Main Article Content
Abstract
Drowsiness in drivers is a critical condition that can compromise road safety, so it is essential to have accurate methods to detect it. This study evaluated the DenseNet121 architecture to detect drowsiness in drivers, focusing on the state of the eyes (open or closed) and the manifestation of yawning. The study design was non-experimental, descriptive and quantitative in approach. A set of 2090 images of drivers, obtained from Kaggle, was used, which were preprocessed with the MediaPipe library to facilitate face detection. The DenseNet121 model achieved an accuracy of 98.46% for the yawning state and 99.62% for the eye state. The confusion matrix showed perfect classifications in both categories. The classification report highlighted an F1-Score, recall and precision of 1.00, evidencing its ability to correctly classify all examples. In real-time testing, the model showed 87% to 97% confidence in yawning and 99% consistency in eyes, albeit with some flaws. These findings highlighted the model's efficiency in detecting drowsiness, suggesting its potential as a valuable tool for road safety.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Los nombres y las direcciones de correo electrónico introducidos en esta revista se usarán exclusivamente para los fines establecidos en ella y no se proporcionarán a terceros o para su uso con otros fines.
How to Cite
References
National Library of Medicine, “Drowsiness,” MedlinePlus Medical Encyclopedia. [Accedido: 20 de septiembre de 2024]. [En línea]. Disponible: https://medlineplus.gov/ency/article/003208.htm
K. Peña Prado, “Somnolencia en conductores de transporte público regular de pasajeros de Lima Metropolitana – Perú,” Tesis de Maestría, Universidad Peruana Cayetano Heredia, Lima, Perú, 2017.
World Health Organization, "Road traffic injuries", 2023. [Accedido: 22 de septiembre de 2024]. [En línea]. Disponible: https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries
Dirección General de Tráfico: “Conducir con sueño o cansancio.” DGT, 2023. [Accedido: 23 de septiembre de 2024]. [En línea]. Disponible: https://www.dgt.es/muevete-con-seguridad/evita-conductas-de-riesgo/Conducir-con-sueno-o-cansancio
Instituto Nacional de Estadística e Informática: “Análisis de los Accidentes de Tránsitos Ocurridos en el Año 2016.”, 2017. [Ac-cedido: 24-sep-2024]. [En línea]. Disponible: https://www.inei.gob.pe/est/lib1528/cap03
Xie and A. Yuille, "Genetic CNN" en Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Baltimore, MD, USA, 2017, pp. 1379-1386.
Y. Zhu y S. Newsam, “DenseNet for dense flow,” en IEEE International Conference on Image Processing (ICIP), 2017, pp. 999-1003, doi: 10.1109/ICIP.2017.8296389.
Ahmed, T., Jyoti, O., Mou, T.H.: Drivers’ drowsiness detection system enhancement using deep learning: CNN-based approach. In: 2023 26th International Conference on Computer and Information Technology (ICCIT), pp. 1–6. IEEE, Cox's Bazar, Bangla-desh (2023). doi: 10.1109/ICCIT60459.2023.10441229.
Namburi, A., Sitpasert, P., Duang-onnam, W.: A CNN-LSTM approach for accurate drowsiness and distraction detection in drivers. ICIC Express Letters 18, 907–917 (2024). doi: 10.24507/icicel.18.09.907.
Lim, B.-E., Ng, K.-W., Ng, S.: Drowsiness detection system through eye and mouth analysis. JOIV: International Journal on Informatics Visualization 7, 2489 (2023). doi: 10.62527/joiv.7.4.2288.
Mohanty, S., Hegde, S.V., Prasad, S., Manikandan, J.: Design of real-time drowsiness detection system using Dlib. In: 2019 IEEE International WIE Conference on Electrical and Computer Engineering (WIECON-ECE), pp. 1–5 (2019). doi: 10.1109/wiecon-ece48653.2019.9019910.
Thampi, L., Kashyap, K., T, N., Reddy, A., Khan, I., Aswathy, M., Kumar, A., Kumar, S.: Smart Driver Assistance: Real-Time Drowsiness Detection Using CNN and Computer Vision. Preprint, Research Square (2024). https://doi.org/10.21203/rs.3.rs-4962655/v1.
Huang, G., Liu, Z., Van Der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4700–4708 (2017).
São Paulo Governo do Estado São Paulo São Todos, “Dirigir com sono causa 20% dos acidentes de trânsito | Governo do Estado de São Paulo”, Governo do Estado de São Paulo, Apr. 15, 2017. [Accedido: 25 de septiembre de 2024]. [En línea]. Disponible: https://www.saopaulo.sp.gov.br/spnoticias/dirigir-com-sono-causa-20-acidentes/.
Dheeraj Perumandla, “Drowsiness_dataset,” Kaggle, 2020. [Accedido: 20 de septiembre de 2024]. [En línea]. Disponible: https://www.kaggle.com/datasets/dheerajperumandla/drowsiness-dataset