Deteksi Postur Duduk Ergonomis Secara Real-Time Menggunakan Pendekatan Hibrida YOLO11n dan Pemetaan Sudut Biomekanik

Faishal Faturrahman; Faiz Alfiansyah; Muhammad Zaidan Athallah Ar Rafif; Rachmat Adi Purnama; Veti Apriana; Sujiliani Heristian

doi:10.58602/jaiti.v4i4.447

Faishal Faturrahman ^{(Corresponding Author)} Universitas Bina Sarana Informatika
Faiz Alfiansyah Universitas Bina Sarana Informatika
Muhammad Zaidan Athallah Ar Rafif Universitas Bina Sarana Informatika
Rachmat Adi Purnama Universitas Bina Sarana Informatika
Veti Apriana Universitas Bina Sarana Informatika
Sujiliani Heristian Universitas Bina Sarana Informatika

DOI: https://doi.org/10.58602/jaiti.v4i4.447

Keywords: Biomekanik, Deteksi Postur Duduk, Ergonomi, Pendekatan Hibrida, YOLO11n

Abstract

Kebiasaan duduk tidak ergonomis di depan komputer dalam durasi berkepanjangan meningkatkan risiko gangguan muskuloskeletal, sementara sistem evaluasi berbasis ekstraksi titik sendi konvensional rentan gagal ketika sebagian anggota tubuh terhalang pakaian maupun objek sekitar. Penelitian ini mengusulkan sistem deteksi postur duduk ergonomis secara real-time melalui pendekatan hibrida yang memadukan arsitektur deep learning YOLO11n dan pemetaan sudut biomekanik berbasis MediaPipe PoseLandmarker menggunakan mekanisme Hybrid Decision Fusion pada aplikasi web SadarPostur. Sebanyak 8.470 citra postur duduk dikurasi ke dalam dua kelas (Good Posture dan Bad Posture) serta dipartisi secara independen dengan rasio 80% latih (6.776 citra), 10% validasi (847 citra), dan 10% uji (847 citra). Model YOLO11n berperan sebagai pengklasifikasi visual utama, sedangkan pemetaan koordinat anatomis telinga, bahu, serta pinggul berfungsi memvisualisasikan garis sumbu tulang belakang sekaligus memverifikasi sudut kemiringan leher dan torso pada kondisi pencahayaan rendah. Pengujian terhadap data uji independen mencatatkan kinerja mAP@50 sebesar 96,41%, mAP@50-95 sebesar 84,90%, presisi makro 92,88%, sensitivitas (recall) makro 87,82%, dan skor F1 makro 90,28% dengan kecepatan inferensi inti 3,9 ms per citra. Capaian tersebut meningkatkan akurasi klasifikasi sebesar 19,02% dibandingkan metode baseline berbasis MediaPipe–Random Forest (77,39%). Sistem diimplementasikan pada antarmuka web interaktif berbasis Streamlit dengan menerapkan kebijakan nir-penyimpanan (Zero-Storage Policy), sehingga mampu menyajikan umpan balik ergonomi mandiri yang cermat, transparan, dan melindungi kerahasiaan visual pengguna.

Downloads

Download data is not yet available.

References

J. Estrada and L. Vea, “Computer Users Sitting Posture Classification Using Distinct Feature Points and Small Scale Convolutional Neural Network for Humana Computer Intelligent Interactive System During COVID-19,” J. Comput. Sci., vol. 19, no. 4, pp. 493–513, 2023, doi: 10.3844/jcssp.2023.493.513.

C. Jandaeng, C. Kongkwamcharoen, T. Jaiwang, P. Koad, and M. Thu, “ANYA: A Noteworthy Youth Annotation Dataset for Machine Learning-Based Ergonomic Posture Analysis,” IEEE Access, vol. 13, pp. 199665–199682, 2025, doi: 10.1109/ACCESS.2025.3633624.

H. COSKUN, “Human Sitting Postures Classification Based on Angular Features with Fuzzy-Logic Labeling,” Int. Conf. Appl. Eng. Nat. Sci., vol. 1, no. 1, pp. 611–619, 2023, doi: 10.59287/icaens.1068.

T. A. Pawitra, F. D. Sitania, A. Septiarini, and H. Hamdani, “Automated ergonomic sitting postures detection for office workstation using XGBoost method,” IAES Int. J. Artif. Intell., vol. 15, no. 1, pp. 506–514, 2026, doi: 10.11591/ijai.v15.i1.pp506-514.

T. A. Pawitra, A. Septiarini, J. I. Tanson, F. D. Sitania, A. E. Burhandenny, and H. Hamdani, “Classification of Ergonomic Sitting Positions Behind a Desk using MobileNetV2,” 2025 9th Int. Conf. Electr. Electron. Inf. Eng. ICEEIE 2025, 2025, doi: 10.1109/ICEEIE66203.2025.11251800.

J. Lee, H. Choi, and J. Kim, “Classification Algorithm for Sitting Postures Using Weighted Random Forest,” IET Image Process., vol. 19, no. 1, Jan. 2025, doi: 10.1049/ipr2.70126.

J. Shen, L. Chen, X. He, C. Zuo, X. Li, and L. Dong, “An Interactive Human-in-the-Loop Framework for Skeleton-Based Posture Recognition in Model Education,” Biomimetics, vol. 10, no. 7, 2025, doi: 10.3390/biomimetics10070431.

S. Chen and Q. Liu, “Efficient Sitting Posture Detection and Sedentary Reminder Application Based on MediaPipe Posture Tracking,” Proc. 2025 3rd Int. Conf. Internet Things Cloud Comput. Technol. IoTCCT 2025, pp. 84–89, 2025, doi: 10.1145/3776865.3776880.

C. Singhtaun, S. Natsupakpong, and P. Lorprasertkul, “Ergonomic Risk Assessment Using Human Pose Estimation with MediaPipe Pose,” ACM Int. Conf. Proceeding Ser., pp. 465–471, 2025, doi: 10.1145/3719384.3719453.

A. Singha et al., “Detection of Wrong Sitting Posture to Prevent Ergonomic Hazards,” Proc. 2026 Int. Conf. Cogn. Comput. Netw. Syst. ICC-CNS 2026, pp. 402–407, 2026, doi: 10.1109/ICC-CNS70518.2026.11606024.

M. Lal, A. Singh, S. A. Ali, and S. Das, “CBRRES: a computer-based REBA and RULA evaluation system for musculoskeletal disorder estimation,” Int. J. Occup. Saf. Ergon., 2026, doi: 10.1080/10803548.2026.2633030.

K. Bourahmoune, K. Ishac, and T. Amagasa, “Intelligent Posture Training: Machine-Learning-Powered Human Sitting Posture Recognition Based on a Pressure-Sensing IoT Cushion,” Sensors, vol. 22, no. 14, 2022, doi: 10.3390/s22145337.

C. Sri-ngernyuang, Prakrankiat Youngkong, Jinpitcha Mamom, and Duangruedee Lasuka, “Sitting Posture Detection and Classification Using Machine Learning Algorithms on RapidMiner,” Emit. Int. J. Eng. Technol., vol. 13, no. 1, pp. 156–172, 2025, doi: 10.24003/emitter.v13i1.898.

T. A. Pawitra et al., “Detection of Ergonomic Sitting Postures in Office Environments,” J. Image Graph. (United Kingdom), vol. 13, no. 6, pp. 686–701, 2025, doi: 10.18178/JOIG.13.6.686-701.

A. Kulikajevas, R. Maskeliunas, and R. Damaševičius, “Detection of sitting posture using hierarchical image composition and deep learning,” PeerJ Comput. Sci., vol. 7, pp. 1–20, 2021, doi: 10.7717/peerj-cs.442.

I. El Bouhali, A. Babur, A. Moukadem, K. Skerl, and A. Dieterlen, “Generalizable features for non-intrusive sitting posture recognition,” Array, vol. 31, p. 101204, 2026, doi: 10.1016/j.array.2026.101204.

R. Sapkota and M. Karkee, “Ultralytics YOLO Evolution: An Overview of YOLO27, YOLO26, YOLO11, YOLOv8, and YOLOv5 Object Detectors for Computer Vision and Pattern Recognition,” Nov. 13, 2026, arXiv. doi: 10.48550/ARXIV.2510.09653.

H. Jin, X. He, L. Wang, Y. Zhu, W. Jiang, and X. Zhou, “SitPose: Real-Time Detection of Sitting Posture and Sedentary Behavior Using Ensemble Learning With Depth Sensor,” IEEE Sens. J., vol. 25, no. 7, pp. 12444–12454, 2025, doi: 10.1109/JSEN.2025.3541821.

P. Nilmart, W. Yodchaisarn, N. Vajidee, N. Tansakul, W. Sramongkhol, and S. Leangsakul, “Immediate effects of sitting position on cervical and trunk posture, perceived discomfort, and RULA-based ergonomic risk during smartphone gaming: A randomized within-subject study,” PLoS One, vol. 21, no. 7, 2026, doi: 10.1371/journal.pone.0354360.

H. Norasi, E. Tetteh, P. Sarker, G. A. Mirka, and M. S. Hallbeck, “Exploring the relationship between neck flexion and neck problems in occupational populations: a systematic review of the literature,” Ergonomics, vol. 65, no. 4, pp. 587–603, 2022, doi: 10.1080/00140139.2021.1976847.

Y. L. Chen and Y. C. Chan, “Neck and shoulder strains under various head flexing positions while standing and sitting with and without back support for male and female smartphone users,” Ergonomics, vol. 67, no. 7, pp. 913–924, 2024, doi: 10.1080/00140139.2023.2270651.