IDENTIFIKASI MENGANTUK MENGGUNAKAN ALGORITMA CNN
DOI:
https://doi.org/10.56127/jukim.v4i2.1946Keywords:
Deteksi Mengantuk, Convolutional Neural Network (CNN), Deep Learning, Keselamatan Berkendara, TensorFlow Lite.Abstract
Kelelahan dan rasa mengantuk merupakan faktor utama yang berkontribusi terhadap kecelakaan lalu lintas dan insiden kerja yang berbahaya. Deteksi kondisi mengantuk pada individu, terutama pengemudi, menjadi hal yang krusial dalam meningkatkan keselamatan. Penelitian ini bertujuan untuk mengembangkan sistem deteksi mengantuk menggunakan Convolutional Neural Network (CNN) berbasis citra wajah. Algoritma CNN memiliki kemampuan unggul dalam mengenali pola visual, termasuk perubahan ekspresi wajah dan pola mata yang menjadi indikator utama kondisi mengantuk.
Penelitian ini menggunakan dataset citra wajah yang mencakup kondisi mata terbuka, mata tertutup, menguap, dan tidak menguap sebagai indikator mengantuk. Model CNN yang dikembangkan dilatih dengan menggunakan teknik preprocessing dan augmentasi data guna meningkatkan akurasi dalam mendeteksi kondisi mengantuk dalam berbagai situasi pencahayaan dan sudut wajah. Model diuji menggunakan metrik evaluasi seperti akurasi, presisi, recall, dan F1-score untuk mengukur efektivitas deteksi.
Hasil penelitian menunjukkan bahwa model CNN yang dikembangkan mampu mendeteksi kondisi mengantuk dengan akurasi mencapai 96% pada dataset uji. Implementasi model menggunakan TensorFlow Lite juga memungkinkan inferensi berjalan lebih cepat dan efisien pada perangkat dengan spesifikasi terbatas. Sistem ini memiliki potensi untuk diterapkan dalam berbagai aplikasi, seperti Driver Monitoring System (DMS) atau sistem pemantauan kerja yang memerlukan deteksi kewaspadaan pengguna.
References
[1] Arduino.cc, 2024. https://docs.arduino.cc/hardware/nicla-vision/ (accessed Mar. 01, 2024).
[2] LeCun, Y., Bengio, Y., & Hinton, G., "Deep learning," Nature, vol. 521, no. 7553, hal. 436-444, 2015.
[3] Krizhevsky, A., Sutskever, I., & Hinton, G. E., "ImageNet Classification with Deep Convolutional Neural Networks," Advances in Neural Information Processing Systems (NeurIPS), 2012.
[4] “Driver Drowsiness Detection With FOMO,” Hackster.io. https://www.hackster.io/coderscafe/driver-drowsiness-detection-with-fomo-206ccf (accessed Mar. 01, 2024).
[5] Z. Zhao, N. Zhou, L. Zhang, H. Yan, Y. Xu, and Z. Zhang, “Driver Fatigue Detection Based on Convolutional Neural Networks Using EM-CNN,” Computational Intelligence and Neuroscience, vol. 2020, p. e7251280, Nov. 2020, doi: https://doi.org/10.1155/2020/7251280.
[6] “Method of Detecting Logistics Driver’s Fatigue State Based on Computer Vision | IEEE Conference Publication | IEEE Xplore,” ieeexplore.ieee.org. https://ieeexplore.ieee.org/abstract/document/7810833/ (accessed Mar. 01, 2024).
[7] P. R. Johns and G. Hocking, "Driver Drowsiness Detection Using Eye Closure Duration," Journal of Sleep Research, vol. 28, no. 5, pp. 567-575, 2019.
[8] in Motor, Committee on National Statistics, Board on Human-Systems Integration, Division of Behavioral and Social Sciences and Education, Transportation Research Board, and National Academies of Sciences, Engineering, and Medicine, “Fatigue, Hours of Service, and Highway Safety,” Nih.gov, Aug. 12, 2016. https://www.ncbi.nlm.nih.gov/books/NBK384974/
[9] Åkerstedt, T., et al., "The Impact of Sleep Quality on Driver Drowsiness: A Literature Review," Journal of Sleep Research, 2004.
[10] Jessica Erin Mabry et al., “Unravelling the Complexity of Irregular Shiftwork, Fatigue and Sleep Health for Commercial Drivers and the Associated Implications for Roadway Safety,” International Journal of Environmental Research and Public Health, vol. 19, no. 22, pp. 14780–14780, Nov. 2022, doi: https://doi.org/10.3390/ijerph192214780.
[11] Dinges, D. F., "PERCLOS: A Valid Psychophysiological Measure of Alertness As Assessed by Psychomotor Vigilance," U.S. Department of Transportation, 1998.
[12] Federal Motor Carrier Safety Administration (FMCSA), "The Validity and Reliability of the PERCLOS Metric," 2016.
[13] Johns, M. W., "A sleep physiologist’s view of the drowsy driver," Transportation Research Part F: Traffic Psychology and Behaviour, vol. 3, no. 4, pp. 241-249, 2000.
[14] A. George dan A. Routray, "Design and Implementation of Real-time Algorithms for Eye Tracking and PERCLOS Measurement for on board Estimation of Alertness of Drivers," arXiv:1505.06162 [cs.CV], May 22, 2015. [Online]. Available: https://doi.org/10.48550/arXiv.1505.06162.
[15] C.-D. Chiu, R. C.-H. Chang, C.-Y. Wang, dan Y.-Y. Kao, "Drowsiness Detection System Based on PERCLOS and Facial Physiological Signal," Sensors, vol. 22, no. 14, p. 5380, Jul. 2022. [Online]. Available: https://doi.org/10.3390/s22145380.
[16] Lv, H., et al., "Driver fatigue detection based on SVM and improved particle swarm optimization," *Neurocomputing*, vol. 173, hal. 1683-1690, 2016.
[17] C. Anderson and J. Horne, "Sleepiness and Driving: The Importance of Early Detection," Traffic Injury Prevention, vol. 18, no. 7, pp. 725-732, 2017.
[18] "Fatigue Detection Based on Eye Movement," Journal of Safety Research, vol. 72, pp. 121-127, 2020.
[19] "Driver Drowsiness Detector System using Raspberry Pi and OpenCV," CircuitDigest. [Online]. Available: https://circuitdigest.com.
[20] "A Review of Recent Developments in Driver Drowsiness Detection Systems," Sensors, MDPI. [Online]. Available: https://www.mdpi.com/1424-8220/19/3/657.
[21] P. Luo, X. Li, Y. Wang, "Driver Drowsiness Detection Based on Eye State Recognition," IEEE Transactions on Human-Machine Systems, vol. 45, no. 4, pp. 574-582, 2021.
[22] "Implementing Real-Time Drowsiness Detection System Using Deep Learning and Raspberry Pi," International Journal of Advanced Computer Science and Applications (IJACSA), 2020.
[23] Raspberry Pi 4 Model B Specifications," Raspberry Pi Foundation. [Online]. Available: https://www.raspberrypi.org/products/raspberry-pi-4-model-b/. [Accessed: 05-Jun-2024].
[24] "Arduino Nicla Vision," Arduino Official Website. [Online]. Available: https://store.arduino.cc/nicla-vision. [Accessed: 05-Jun-2024].
[25] D. Molloy, Exploring Raspberry Pi: Interfacing to the Real World with Embedded Linux, Wiley, 2016.
[26] S. Monk, Programming Arduino: Getting Started with Sketches, 2nd ed., McGraw-Hill Education TAB, 2016.
[27] J. B. Polsonetti, "How to Power Your Raspberry Pi: Power Requirements, Performance, and Safety," Electronic Design, 2019. [Online]. Available: https://www.electronicdesign.com/technologies/embedded-revolution/article/21121215/how-to-power-your-raspberry-pi-power-requirements-performance-and-safety. [Accessed: 05-Jun-2024].
[28] "Arduino Nicla Vision Datasheet," Arduino Documentation, 2020. [Online]. Available: https://docs.arduino.cc/resources/datasheets/Nicla-Vision-datasheet.pdf. [Accessed: 05-Jun-2024].
[29] Zhao, Z., et al. (2020). “EM-CNN dan YOLO untuk deteksi kelelahan.” Jurnal Internasional.
[30] Luo, P., Li, X., & Wang, Y. (2021). “Pengakuan status mata menggunakan YOLO.” Jurnal Internasional.
[31] Chiu, C.-D., et al. (2022). “PERCLOS dan YOLO untuk deteksi kantuk.” Jurnal Internasional.
[32] Mabry, J. E., et al. (2022). Analisis Shiftwork dan Kelelahan. Jurnal Internasional.
[33] Driver Fatigue Detection Based on EM-CNN (2023). EM-CNN untuk deteksi kantuk. Jurnal Internasional.
[34] Luo, et al. (2023). Pendeteksian kantuk dengan algoritma YOLO. Jurnal Internasional.
[35] Journal of Safety Research (2024). Deteksi kantuk berbasis gerakan mata. Jurnal Internasional.
[36] Johns, P. R., & Hocking, G. (2024). Pengukuran durasi mata tertutup. Jurnal Internasional.
