Human Body Posture Classification from Digital Images Using the InceptionV3 Architecture
DOI:
https://doi.org/10.30871/jaic.v10i4.13571Keywords:
Human Body Posture, Image Classification, InceptionV3, Transfer Learning, MPII Human Pose DatasetAbstract
This study aims to develop a human body posture classification model based on digital images using the InceptionV3 architecture. The dataset used in this study is the MPII Human Pose Dataset, which contains a wide variety of human activities and body postures. The research began with label extraction from a metadata file in .mat format, followed by the selection of the 20 activity classes with the largest number of samples. Subsequently, the dataset was balanced using an undersampling technique, resulting in 140 images for each class. The images were then resized to 224×224 pixels, normalized using the preprocess_input function, and enhanced through data augmentation applied to the training set. The model was developed using a transfer learning approach with InceptionV3 as the base model. Additional layers, including Global Average Pooling, Dropout, and Dense layers, were added to perform multi-class classification. The experimental results showed that the proposed model achieved a test accuracy of 0.8893 with a test loss of 0.5587. Furthermore, the macro-average metrics obtained from the classification report were 0.9018 for precision, 0.8893 for recall, and 0.8877 for F1-score. These results indicate that the model was able to classify most activity classes effectively. However, several classes with similar visual characteristics still caused misclassification, indicating opportunities for further improvement in human posture recognition performance.
Downloads
References
[1] V. T. Le, K. Tran-Trung, and V. T. Hoang, “A Comprehensive Review of Recent Deep Learning Techniques for Human Activity Recognition,” Comput. Intell. Neurosci., vol. 2022, 2022, doi: 10.1155/2022/8323962.
[2] S. Zhang et al., “Deep Learning in Human Activity Recognition withWearable Sensors: A Review on Advances,” Sensors, vol. 22, no. 4, pp. 1–43, Feb. 2022, doi: 10.3390/s22041476.
[3] 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, Mar. 2021, doi: 10.7717/peerj-cs.442.
[4] M. D. Meitantya, C. A. Sari, E. H. Rachmawanto, and R. R. Ali, “VGG-16 ARCHITECTURE ON CNN FOR AMERICAN SIGN LANGUAGE CLASSIFICATION,” Jurnal Teknik Informatika (Jutif), vol. 5, no. 4, pp. 1165–1171, Jul. 2024, doi: 10.52436/1.jutif.2024.5.4.2160.
[5] R. O. Ogundokun, R. Maskeliūnas, and R. Damaševičius, “Human Posture Detection Using Image Augmentation and Hyperparameter-Optimized Transfer Learning Algorithms,” Applied Sciences (Switzerland), vol. 12, no. 19, pp. 1–23, Oct. 2022, doi: 10.3390/app121910156.
[6] A. R. Rababaah, “Deep Learning of Human Posture Image Classification using Convolutional Neural Networks,” International Journal of Computer Science and Mathematics, vol. X, pp. 1–11.
[7] A. Susanto, Y. Kusumawati, E. Dhimas Niagara, and C. A. Sari, “CONVOLUTIONAL NEURAL NETWORK DALAM SISTEM DETEKSI HELM PADA PENGENDARA MOTOR,” 2 st Proceeding STEKOM, vol. 2022, 2022.
[8] M. Mesárošová, O. Mihálik, and M. Jirgl, “CNN Architecture for Posture Classification on Small Data,” in IFAC-PapersOnLine, Elsevier B.V., Jun. 2024, pp. 299–304. doi: 10.1016/j.ifacol.2024.07.413.
[9] L. Zhang, L. Zhao, and Y. Yan, “A hybrid neural network-based intelligent body posture estimation system in sports scenes,” Mathematical Biosciences and Engineering, vol. 21, no. 1, pp. 1017–1037, 2024, doi: 10.3934/mbe.2024042.
[10] V. A. Aydın, “Comparison of CNN-based methods for yoga pose classification,” Turkish Journal of Engineering, vol. 8, no. 1, pp. 65–75, 2024, doi: 10.31127/tuje.1275826.
[11] F. Bin Ashraf, M. U. Islam, M. R. Kabir, and J. Uddin, “YoNet: A Neural Network for Yoga Pose Classification,” SN Comput. Sci., vol. 4, no. 2, pp. 1–9, Mar. 2023, doi: 10.1007/s42979-022-01618-8.
[12] R. O. Ogundokun, R. Maskeliūnas, S. Misra, and R. Damasevicius, “Hybrid InceptionV3-SVM-Based Approach for Human Posture Detection in Health Monitoring Systems,” Algorithms, vol. 15, no. 11, Nov. 2022, doi: 10.3390/a15110410.
[13] H. C. Nguyen, T. H. Nguyen, R. Scherer, and V. H. Le, “Deep Learning for Human Activity Recognition on 3D Human Skeleton: Survey and Comparative Study,” Sensors, vol. 23, no. 11, pp. 1–33, Jun. 2023, doi: 10.3390/s23115121.
[14] E. Samkari, M. Arif, M. Alghamdi, and M. A. Al Ghamdi, “Human Pose Estimation Using Deep Learning: A Systematic Literature Review,” Mach. Learn. Knowl. Extr., vol. 5, no. 4, pp. 1612–1659, Dec. 2023, doi: 10.3390/make5040081.
[15] R. O. Ogundokun, R. Maskeliūnas, S. Misra, and R. Damasevicius, “A Novel Deep Transfer Learning Approach Based on Depth-Wise Separable CNN for Human Posture Detection,” Information (Switzerland), vol. 13, no. 11, pp. 1–22, Nov. 2022, doi: 10.3390/info13110520.
[16] A. A. Istiqomah, C. A. Sari, A. Susanto, and H. Rachmawanto, “Facial Expression Recognition using Convolutional Neural Networks with Transfer Learning Resnet-50,” 2024. [Online]. Available: http://jurnal.polibatam.ac.id/index.php/JAIC
[17] I. Kareem, S. F. Ali, M. Bilal, and M. S. Hanif, “Exploiting the features of deep residual network with SVM classifier for human posture recognition,” PLoS One, vol. 19, no. 12 December, pp. 1–26, Dec. 2024, doi: 10.1371/journal.pone.0314959.
[18] L. Stern and A. R. Fekr, “In-Bed Posture Classification Using Deep Neural Network,” Sensors, vol. 23, no. 5, pp. 1–10, Mar. 2023, doi: 10.3390/s23052430.
[19] H. Chen and R. Fan, “Improved convolutional neural network for precise exercise posture recognition and intelligent health indicator prediction,” Sci. Rep., vol. 15, no. 1, pp. 1–16, Dec. 2025, doi: 10.1038/s41598-025-01854-x.
[20] B. A. Khan and J. W. Jung, “Deep Learning-Based Human Activity Recognition Using Dilated CNN and LSTM on Video Sequences of Various Actions Dataset,” Applied Sciences (Switzerland), vol. 15, no. 22, pp. 1–18, Nov. 2025, doi: 10.3390/app152212173.
[21] G. S. Bang and S. B. Park, “Workout Classification Using a Convolutional Neural Network in Ensemble Learning,” Sensors, vol. 24, no. 10, pp. 1–15, May 2024, doi: 10.3390/s24103133.
[22] A. S. Talaat, “Novel deep learning models for yoga pose estimator,” SN Appl. Sci., vol. 5, no. 12, pp. 1–13, Dec. 2023, doi: 10.1007/s42452-023-05581-8.
[23] A. Iazzi, M. Rziza, and R. O. H. Thami, “Fall detection system-based posture-recognition for indoor environments,” J. Imaging, vol. 7, no. 3, pp. 1–24, Mar. 2021, doi: 10.3390/jimaging7030042.
[24] 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,” Journal of Computer Science, vol. 19, no. 4, pp. 493–513, 2023, doi: 10.3844/jcssp.2023.493.513.
[25] H. Yoo and K. Chung, “Classification of Multi-Frame Human Motion Using CNN-based Skeleton Extraction,” Intelligent Automation and Soft Computing, vol. 34, no. 1, pp. 1–13, 2022, doi: 10.32604/iasc.2022.024890.
[26] M. H. de la Cruz, U. Solache, A. Luna-Álvarez, S. R. Zagal-Barrera, D. A. M. López, and D. Mujica-Vargas, “CNN 1D: A Robust Model for Human Pose Estimation,” Information (Switzerland), vol. 16, no. 2, pp. 1–20, Feb. 2025, doi: 10.3390/info16020129.
[27] X. Cao, X. Wang, X. Geng, D. Wu, and H. An, “An Approach for Human Posture Recognition Based on the Fusion PSE-CNN-BiGRU Model,” CMES - Computer Modeling in Engineering and Sciences, vol. 140, no. 1, pp. 385–408, 2024, doi: 10.32604/cmes.2024.046752.
[28] Z. Xu, W. Qu, H. Cao, M. Dong, D. Li, and Z. Qiu, “An Adaptive Human Posture Detection Algorithm Based on Generative Adversarial Network,” Comput. Intell. Neurosci., vol. 2022, pp. 1–9, 2022, doi: 10.1155/2022/7193234.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Munandar Rahmat Prayogi, Christy Atika Sari, Eko Hari Rachmawanto
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:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement 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 acknowledgement 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 (See The Effect of Open Access).
