Face Recognition Using MTCNN Face Detection, ResNetV1 Feature Embeddings, and SVM Classification
DOI:
https://doi.org/10.30871/jaic.v9i5.11016Keywords:
Face Recognition, MTCNN, ResNetV1, Support Vector Machine, Deep LearningAbstract
Face recognition has become an essential component of modern security and authentication systems, yet its effectiveness is often challenged by limited datasets, class imbalance, variations in facial poses, lighting conditions, and image resolutions. This study proposes a face recognition pipeline that integrates Multi-task Cascaded Convolutional Networks (MTCNN) for face detection, Residual Network V1 (ResNetV1) for feature extraction, and Support Vector Machine (SVM) for classification. Unlike previous works that rely on large-scale datasets and end-to-end deep learning models, this study emphasizes the effectiveness of the pipeline under constrained data conditions, using 856 images across 191 classes with highly imbalanced distribution. Experimental results show that MTCNN successfully detected 97.1% of faces, while ResNetV1 produced 512-dimensional embeddings that formed well-separated clusters validated by clustering metrics (Silhouette Score = 0.578, Davies-Bouldin Index = 0.566). The SVM classifier achieved 92.9% accuracy, with macro-average precision, recall, and F1-scores of 0.89, 0.92, and 0.89 respectively, significantly outperforming a baseline k-Nearest Neighbor (k-NN) model that only reached 63.9% accuracy. These findings highlight the novelty of this study: demonstrating that a lightweight yet robust pipeline can deliver reliable recognition performance even in small, imbalanced datasets, making it suitable for real-world scenarios where large-scale training data are not available.
Downloads
References
[1] B. Hartika and D. Ahmad, "Face Recognition Menggunakan Algoritma Haar Cascade Classifier dan Convolutional Neural Network," Journal of Mathematics UNP, vol. 2, no. 1, pp. 1–7, 2022. [Online]. Available: https://ejournal.unp.ac.id/students/index.php/mat/article/view/11954
[2] O. A. Naser, S. Mumtazah, K. Samsudin, M. Hanafi, S. M. Binti, and N. Z. Zamri, "Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions," Journal of Communications Software and Systems, vol. 21, no. 1, pp. 109–119, Mar. 2025, doi: 10.24138/jcomss-2024-0084.
[3] D. M. Abdulhussien and L. J. Saud, “Evaluation Study of Face Detection by Viola-Jones Algorithm,” International journal of health sciences, pp. 4174–4182, Sep. 2022, doi: https://doi.org/10.53730/ijhs.v6ns8.13127.
[4] F. R. Chandra, A. Nur, and R. Hidayat, "Analysis of the use of MTCNN and landmark technology to improve the accuracy of facial recognition on official documents," Journal of Applied Informatics and Computing, vol. 9, no. 2, pp. 112–120, 2025. [Online]. Available: https://jurnal.polibatam.ac.id/index.php/JAIC/article/view/8814
[5] Z. Li et al., “A classification method for multi-class skin damage images combining quantum computing and Inception-ResNet-V1,” Frontiers in Physics, vol. 10, Nov. 2022, doi: https://doi.org/10.3389/fphy.2022.1046314.
[6] S. Almabdy and L. Elrefaei, “Deep Convolutional Neural Network-Based Approaches for Face Recognition,” Applied Sciences, vol. 9, no. 20, p. 4397, Oct. 2019, doi: https://doi.org/10.3390/app9204397.
[7] R. E. Saragih and Q. H. To, “A Survey of Face Recognition based on Convolutional Neural Network,” Indonesian Journal of Information Systems, vol. 4, no. 2, Feb. 2022, doi: https://doi.org/10.24002/ijis.v4i2.5439.
[8] P. P. Raj, “An Evaluation of MTCNN in Face Recognition Algorithms for Effective Detection in Masked Scenarios in Real Time Video Surveillance,” African Journal of Biomedical Research, pp. 12252–12261, Dec. 2024, doi: https://doi.org/10.53555/ajbr.v27i4s.6154.
[9] M. Yuan, Seyed Yahya Nikouei, Alem Fitwi, Y. Chen, and Y. Dong, “Minor Privacy Protection Through Real-time Video Processing at the Edge,” arXiv (Cornell University), Aug. 2020, doi: https://doi.org/10.1109/icccn49398.2020.9209632.
[10] “A. I. Awodeyi, O. A. Ibok, I. Omokaro, J. U. Ekwemuka, and M. O. Ighofiomoni, “Effective preprocessing techniques for improved facial recognition under variable conditions,” Franklin Open, vol. 10, p. 100225, Jan. 2025, doi: https://doi.org/10.1016/j.fraope.2025.100225.
[11] A. Jan, S. Abid, M. F. Khan, A. Hussain, and A. A. Khuhro, “Evaluation of Pre-Trained CNN Models for Face Recognition,” Sensors, vol. 23, no. 6, p. 2901, 2023. [Online]. Available: https://doi.org/10.3390/s23062901
[12] I. C. Duta, L. Liu, F. Zhu, and L. Shao, "Improved Residual Networks for Image and Video Recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 43, no. 12, pp. 4135–4147, Dec. 2021, doi: https://doi.org/10.1109/TPAMI.2021.3055457
[13] Y. Yang et al., “A comparative analysis of eleven neural networks architectures for small datasets of lung images of COVID-19 patients toward improved clinical decisions,” Computers in Biology and Medicine, vol. 139, p. 104887, Dec. 2021, doi: https://doi.org/10.1016/j.compbiomed.2021.104887.
[14] P. Hofer, M. Roland, P. Schwarz, and R. Mayrhofer, "Shrinking embeddings, not accuracy: Performance-preserving reduction of facial embeddings for complex face verification computations," Johannes Kepler Univ. Linz, Austria, Tech. Rep., 2023. [Online]. Available: https://www.researchgate.net/publication/384278329
[15] H. Zhang, “Real-time face recognition method based on MTCNN-Inception-ResNet-v2-SVM model,” Applied and Computational Engineering, vol. 45, no. 1, pp. 179–189, Mar. 2024, doi: https://doi.org/10.54254/2755-2721/45/20241677.
[16] J. Opitz, “A Closer Look at Classification Evaluation Metrics and a Critical Reflection of Common Evaluation Practice,” Transactions of the Association for Computational Linguistics, vol. 12, pp. 820–836, Jan. 2024, doi: https://doi.org/10.1162/tacl_a_00675.
[17] G. Vardakas, I. Papakostas, and A. Likas, "Deep clustering using the soft silhouette score: Towards compact and well-separated clusters," in *Proceedings of the International Joint Conference on Neural Networks (IJCNN)*, 2020, pp. 1–8.
[18] Yuli Asriningtias and Joko Aryanto, “K-Means Algorithm with Davies Bouldin Criteria for Clustering Provinces in Indonesia Based on Number of Events and Impacts of Natural Disasters,” International Journal of Engineering Technology and Natural Sciences, vol. 4, no. 1, pp. 75–80, Jul. 2022, doi: https://doi.org/10.46923/ijets.v4i1.147.
[19] Z. Syahputri, Sutarman, dan M. A. P. Siregar, “Determining The Optimal Number of K-Means Clusters Using The Calinski Harabasz Index and Krzanowski and Lai Index Methods for Grouping Flood Prone Areas In North Sumatra,” SinkrOn, vol. 8, no. 1, pp. 571–580, Jan. 2024.
[20] H. Tariq, M. Majeed, and M. Ahmad, "Optimizing SVM Performance through Combinatorial Hyperparameter Tuning and Model Selection," Univ. of Agriculture Faisalabad, Pakistan, 2025. [Online]. Available: https://www.researchgate.net/publication/393104826_Optimizing_SVM_Performance_through_Combinatorial_Hyperparameter_Tuning_and_Model_Selection
[21] N. R. Feta, “Comparison of KNN and SVM Algorithms in Facial Image Recognition Using Haar-Wavelet Feature Extraction,” Information Systems and Technology, Indonesia Cyber University, 2023. [Online]. Available: https://www.researchgate.net/publication/371876569_Comparison_of_KNN_and_SVM_Algorithms_in_Facial_Image_Recognition_Using_Haar_Wavelet_Feature_Extraction. Accessed: Sep. 8, 2025.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Ivan Putra Pratama, Novita Kurnia Ningrum
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).
