Driver Drowsiness Detection Using LSTM and MediaPipe FaceMesh-Based Facial Landmark Analysis
Keywords:
Drowsiness detection, Driver, Long Short-Term Memory (LSTM), MediaPipe FaceMesh, Face classification, Deep learning, Video analysisAbstract
This research is motivated by the high number of traffic accidents caused by human factors, particularly driver drowsiness, which is often not detected early. This condition has the potential to reduce concentration and increase the risk of accidents. Therefore, this study aims to develop a facial classification model to automatically detect driver drowsiness based on deep learning technology. The method used is Long Short-Term Memory (LSTM) to process sequential video data, and MediaPipe FaceMesh to extract facial features consisting of 478 landmark points. The dataset consists of 220 videos with two classes: drowsy and non-drowsy, divided into training, validation, and testing data. The results show that the proposed model achieved an accuracy of 95.5%, with a precision of 95.9%, a recall of 95.5%, and an F1-score of 95.5%. This indicates that the model performs well in distinguishing between drowsy and non-drowsy states. In conclusion, the combination of LSTM and MediaPipe is effective for video-based sleepy facial classification. The contribution of this research is the development of an accurate drowsiness detection model that has the potential to be developed into a real-time driver monitoring system to improve driving safety.
References
[1] M.-Z. Liu, X. Xu, J. Hu, and Q.-N. Jiang, “Real time detection of driver fatigue based on CNN-LSTM,” IET Image Process., vol. 16, no. 2, pp. 576–595, 2022.
[2] L. Yu, X. Yang, H. Wei, J. Liu, and B. Li, “Driver fatigue detection using PPG signal, facial features, head postures with an LSTM model,” Heliyon, vol. 10, no. 21, 2024.
[3] S. A. El-Nabi, W. El-Shafai, E.-S. M. El-Rabaie, K. F. Ramadan, F. E. Abd El-Samie, and S. Mohsen, “Machine learning and deep learning techniques for driver fatigue and drowsiness detection: a review,” Multimed. Tools Appl., vol. 83, no. 3, pp. 9441–9477, 2024, doi: https://doi.org/10.1007/s11042-023-15054-0 .
[4] T. Fonseca and S. Ferreira, “Drowsiness Detection in Drivers: A Systematic Review of Deep Learning-Based Models,” Appl. Sci., vol. 15, p. 9018, Aug. 2025, doi: 10.3390/app15169018.
[5] Z. Gomolka, D. Kordos, E. Dudek-Dyduch, and B. Twarog, “New Perspectives on Eye-Tracking: Theory, Methods, and Applications,” Appl. Sci., vol. 15, no. 21, p. 11463, 2025, doi: https://doi.org/10.3390/app152111463.
[6] S. Bendjebar, Y. Lafifi, R. Boudjehem, and A. Laouissi, “Data-Driven Insights into E-Learning: A Comprehensive Review of Eye-Tracking Applications in Learning Systems,” J. Eye Mov. Res., vol. 19, no. 2, p. 41, 2026, doi: 10.3390/jemr19020041.
[7] F. Nemsia, Y. Zouhir, M. Zarka, and K. Ouni, “Improving Object Detection and Tracking: A Hybrid Method for Real-Time Applications,” in 2025 IEEE International Conference on Advanced Systems and Emergent Technologies (IC_ASET), IEEE, 2025, pp. 1–6. doi: DOI: 10.1109/IC_ASET65966.2025.11232214.
[8] V. V Arlazarov, D. P. Matalov, D. P. Nikolaev, and S. A. Usilin, “Evolution of the Viola-Jones object detection method: A survey,” Bull. South Ural State Univ. Ser. Math. Model. Program. Comput. Softw., vol. 14, no. 4, pp. 5–23, 2021, doi: DOI: 10.14529/mmp210401.
[9] L. Chen, G. Xin, Y. Liu, and J. Huang, “Driver fatigue detection based on facial key points and LSTM,” Secur. Commun. Networks, vol. 2021, no. 1, p. 5383573, 2021, doi: https://doi.org/10.1155/2021/5383573.
[10] W. Utomo, Y. Suhanda, H. Ar-Rasyid, and A. Dharmalau, “Indonesian Language Sign Detection using Mediapipe with Long Short-Term Memory (LSTM) Algorithm,” J. Informatics Web Eng., vol. 4, no. 3, pp. 245–258, 2025, doi: 10.33093/jiwe.2025.4.3.15.
[11] J. F. Torres, F. Martínez-Álvarez, and A. Troncoso, “A deep LSTM network for the Spanish electricity consumption forecasting,” Neural Comput. Appl., vol. 34, pp. 10533–10545, 2022, doi: 10.1007/s00521-021-06773-2.
[12] N. Akhtar, J. Fan, A. R. Buzdar, M. Ahmed, and A. Raza, “VLSI Design of LSTM-Based ECG Classification for Continuous Cardiac Monitoring on Wearable Devices,” Electron. Lett., vol. 61, p. e70269, 2025, doi: 10.1049/ell2.70269.
[13] M. I. B. Ahmed et al., “A Deep-Learning Approach to Driver Drowsiness Detection,” Safety, vol. 9, no. 3, p. 65, 2023, doi: 10.3390/safety9030065.
[14] G. Nguyen, L. Wang, Y. Jiang, and T. Gedeon, “Detecting Fake News Belief via Skin and Blood Flow Signals,” IEEE Access, vol. 14, pp. 30265–30277, 2026, doi: 10.1109/ACCESS.2024.3423723.
[15] E. Essel, F. Lacy, F. Albalooshi, W. Elmedany, and Y. Ismail, “Drowsiness Detection in Drivers Using Facial Feature Analysis,” Appl. Sci., vol. 15, no. 1, p. 20, 2025, doi: 10.3390/app15010020.
[16] O. O. Ajayi, A. M. Kurien, K. Djouani, and L. Dieng, “A Multimodal Systematic Review of Drivers’ Fatigue Detection Methodologies, Datasets, and Models,” IEEE Access, vol. 13, pp. 158266–158284, 2025, doi: 10.1109/ACCESS.2025.3606900.
[17] Y. Albadawi, A. AlRedhaei, and M. Takruri, “Real-Time Machine Learning-Based Driver Drowsiness Detection Using Visual Features,” J. Imaging, vol. 9, p. 91, 2023, doi: 10.3390/jimaging9050091.
[18] X. Jiao, “Multi-Fatigue Indicator Detection for Drivers Using FaceMesh,” in 2025 6th International Conference on Electronic Communication and Artificial Intelligence (ICECAI), 2025, pp. 396–399. doi: 10.1109/icecai66283.2025.11171211.
[19] F. Badri, S. U. R. Sari, and S. A. Bin Hamzah, “Analysis of Driver Drowsiness Detection System Based on Landmarks and MediaPipe,” Inf. J. Ilm. Bid. Teknol. Inf. dan Komun., vol. 10, no. 1, 2025, doi: 10.25139/inform.v10i1.9325.
[20] B. Akrout and S. Fakhfakh, “How to Prevent Drivers before Their Sleepiness Using Deep Learning-Based Approach,” Electronics, vol. 12, p. 965, 2023, doi: 10.3390/electronics12040965.
[21] T. Gautam, “Novel CNN Based Model Using LSTM for Driver Drowsiness Detection,” Int. J. Res. Appl. Sci. Eng. Technol., vol. 13, no. 4, 2025, doi: 10.22214/ijraset.2025.69131.
[22] S. Das, “IoT-Assisted Automatic Driver Drowsiness Detection through Facial Movement Analysis Using Deep Learning and a U-Net-Based Architecture,” Information, vol. 15, p. 30, 2024, doi: 10.3390/info15010030.
[23] S. Dalve, I. Ramdasi, G. Kothawade, Y. Khadke, and M. Wete, “Real Time Prevention of Driver Fatigue Using Deep Learning and MediaPipe,” Int. J. Innov. Res. Comput. Sci. Technol., vol. 11, no. 3, pp. 7–11, 2023, doi: 10.55524/ijircst.2023.11.3.2.
[24] S. Rathod, T. Mali, Y. Jogani, N. Faldu, V. Odedra, and P. Barik, “RealD3: A Real-time Driver Drowsiness Detection Scheme Using Machine Learning,” in 2023 IEEE Wireless Antenna and Microwave Symposium (WAMS), 2023. doi: 10.1109/wams57261.2023.10242860.
