Pengembangan Sistem Deteksi Kantuk Real-Time Berbasis Komputasi Ringan Menggunakan Analisis Sekuensial Geometri Wajah
DOI:
https://doi.org/10.29407/cf83a826Abstract
Kecelakaan lalu lintas yang diakibatkan oleh penurunan kewaspadaan pengemudi, khususnya fenomena microsleep, masih menjadi penyumbang utama fatalitas di jalan raya. Pengembangan sistem pemantauan pengemudi secara non-intrusif menggunakan visi komputer telah banyak dilakukan, namun sebagian besar model deep learning membutuhkan daya komputasi yang besar. Penelitian awal ini bertujuan untuk mengevaluasi kelayakan arsitektur Long Short-Term Memory (LSTM) berbobot ringan untuk mendeteksi kondisi kantuk secara real-time. Pendekatan yang diusulkan mengekstraksi jarak geometri pada area mata (Eye Aspect Ratio) dan mulut (Mouth Aspect Ratio) menggunakan modul MediaPipe Face Mesh. Untuk mengurangi beban komputasi, data deret waktu dibatasi hanya pada 10 sekuens frame per deteksi dan diproses menggunakan arsitektur LSTM tunggal (single-layer). Pengujian dilakukan menggunakan skema train-test split dengan rasio 80:20. Hasil pengujian menunjukkan bahwa model komputasi ringan ini mampu mencapai tingkat akurasi pengujian sebesar 70.83% dalam mengklasifikasikan kondisi terjaga dan mengantuk. Hal ini mengindikasikan bahwa penggunaan sekuens pendek berbasis landmark geometri wajah sangat layak dan berpotensi untuk diintegrasikan pada perangkat dengan sumber daya terbatas sebagai landasan pengembangan model yang lebih kompleks.
Keywords:
Deteksi Kantuk, Microsleep, MediaPipe, LSTM, Computer Vision##plugins.themes.default.displayStats.downloads##
References
[1] M. F. Aslan, A. Durdu, dan K. Sabanci, "Driver Fatigue Detection Based on Eye State Using Deep Learning and Face Mesh," IEEE Access, vol. 9, hlm. 115284-115294, 2021, doi: 10.1109/ACCESS.2021.3105578.
[2] E. Magán, M. P. Sesmero, J. M. Alonso-Weber, dan A. Sanchis, "Driver Drowsiness Detection by Applying Deep Learning Techniques to Sequences of Images," Applied Sciences (Switzerland), vol. 12, no. 3, Feb 2022, doi: 10.3390/app12031145.
[3] R. G. Cinisli, A. Erciyas, dan T. E. Eray, "Real-Time Driver Drowsiness Detection Using Facial Landmarks and Deep Learning," Electronics (Basel), vol. 12, no. 1, hlm. 154, 2023, doi: 10.3390/electronics12010154.
[4] Z. Chen, J. Wu, Y. Shen, dan S. Li, "Driver Drowsiness Detection Based on Spatiotemporal Representation Learning," IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 2, hlm. 1823-1834, 2023, doi: 10.1109/TITS.2022.3218675.
[5] S. Priyanka, S. Shanthi, A. Saran Kumar, dan V. Praveen, "Data fusion for driver drowsiness recognition: A multimodal perspective," Egyptian Informatics Journal, vol. 27, Sep 2024, doi: 10.1016/j.eij.2024.100529.
[6] C. Xu, W. Huang, J. Liu, dan L. Li, "Detecting Driver Drowsiness Using Hybrid Facial Features and Ensemble Learning," Information (Switzerland), vol. 16, no. 4, Apr 2025, doi: 10.3390/info16040294.
[7] A. A. Alawami, S. Almasri, dan A. Alghamdi, "A Lightweight Deep Learning Framework for Driver Drowsiness Detection," Applied Sciences, vol. 13, no. 6, hlm. 3846, 2023, doi: 10.3390/app13063846.
[8] S. A. Almutairi dan G. B. Suliman, "Real-Time Detection of Driver Drowsiness Using Machine Learning Techniques," Information, vol. 14, no. 9, hlm. 504, 2023, doi: 10.3390/info14090504.
[9] Y. Rahmawati, K. Woraratpanya, I. Ardiyanto, dan H. A. Nugroho, "Evaluating the effectiveness of facial actions features for the early detection of driver drowsiness in driving safety monitoring system," Communications in Science and Technology, vol. 10, no. 1, hlm. 179-189, 2025, doi: 10.21924/cst.10.1.2025.1594.
[10] H. Zhang, "Driver Drowsiness Monitoring System Based on Facial Landmarks and Embedded Machine Learning Model," Journal of Intelligent Transportation and Safety, vol. 5, no. 2, hlm. 88-97, 2022, doi: 10.5664/jits.2022.0288.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Copyright on any article is retained by the author(s).
- The author grants the journal, right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work’s authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal’s published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
- The article and any associated published material is distributed under the Creative Commons Attribution-ShareAlike 4.0 International License
