Klasifikasi Bentuk Wajah Menggunakan Efficientnet-B4

Authors

  • John Christofel Wicaksono Universitas Nusantara PGRI Kediri Indonesia
  • Julian Sahertian Universitas Nusantara PGRI Kediri Indonesia
  • Rony Heri Irawan Universitas Nusantara PGRI Kediri Indonesia

DOI:

https://doi.org/10.29407/7dfa6t91

Abstract

Klasifikasi bentuk wajah merupakan komponen penting dalam sistem rekomendasi produk personalisasi seperti kacamata atau kosmetik. Penelitian ini  bertujuan untuk mengembangkan sistem ini bertujuan untuk mengembangkan sistem klasifikasi bentuk wajah otomatis menggunakan srditektur EffiientNet-B4 dengan pendekatan transfer learning pada framework PyTorch. Dataset yang digunakan terdiri dari lima kelas bentuk wajah yaitu oval, round, heart, square dan oblong, yang diambil dari dataset publik. Model dilatih menggunakan augmentasi data, normalisasi, mixed precision training, dan scheduler learning rate. Hasil evaluasi menunjukan bahwa model berhasil mencapai akurasi validasi sebesar 81% dan F1-score rata-rata yang tinggi pada seluruh kelas. Studi ini menunjukan bahwa arsitektur EfficientNet-B4 efektis digunakan dalam tugas klasifikasi bentuk wajah dan dapat digunakan sebagai dasar untuk sistem rekomendasi sebagai dasar untuk sistem rekomendasi berbasis wajah dimasa depan.

Keywords:

deep learning, EfficientNet-B4, klasifikasi wajah, PyTorch, transfer learning

##plugins.themes.default.displayStats.downloads##

##plugins.themes.default.displayStats.noStats##

References

[1] Y. Taigman, M. A. Ranzato, T. Aviv, and M. Park, “Taigman_DeepFace_Closing_the_2014_CVPR_paper”, doi: 10.1109/CVPR.2014.220.

[2] S. Pokhrel, “SISTEM REKOMENDASI FRAME KACAMATA BERDASARKAN BENTUK WAJAH MENGGUNAKAN DEEP LEARNING DENGAN METODE COVOLUTIONAL NEURAL NETWORK ARSITEKTUR RESNET-50,” Αγαη, vol. 15, no. 1, pp. 37–48, 2024, [Online]. Available: https://repository.unissula.ac.id/37527/1/Teknik Informatika_32602000033_fullpdf.pdf

[3] S. Young, F. Natalia, S. Sudirman, and C. S. Ko, “Eyeglasses frame selection based on oval face shape using convolutional neural network,” ICIC Express Lett. Part B Appl., vol. 10, no. 8, pp. 707–715, 2019, doi: 10.24507/icicelb.10.08.707.

[4] N. Dewi and F. Ismawan, “Implementasi Deep Learning Menggunakan Cnn Untuk Sistem Pengenalan Wajah,” Fakt. Exacta, vol. 14, no. 1, p. 34, 2021, doi: 10.30998/faktorexacta.v14i1.8989.

[5] M. E. Prasetyo, M. R. Faza, R. Pratama, S. N. H. Alhabsy, H. Purwanti, and A. P. A. Masa, “Klasifikasi Ragam Kendaraan Menggunakan Metode Convolutional Neural Network (Cnn),” Adopsi Teknol. dan Sist. Inf., vol. 2, no. 2, pp. 142–148, 2023, doi: 10.30872/atasi.v2i2.1156.

[6] R. Adityatama and A. T. Putra, “Image classification of Human Face Shapes Using Convolutional Neural Network Xception Architecture with Transfer Learning,” Recursive J. Informatics, vol. 1, no. 2, pp. 102–109, 2023, doi: 10.15294/rji.v1i2.70774.

[7] Q. V. Le Mingxing Tan, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks Mingxing,” Can. J. Emerg. Med., vol. 15, no. 3, p. 190, 2019.

[8] R. Adolph, “Penerapan Efficient-Net Dalam Mengklasifikasikan Kanker Kulit,” Penerapan Effic. Dalam Mengklasifikasikan Kanker Kulit, pp. 1–23, 2016, [Online]. Available: https://jurnal.unprimdn.ac.id/index.php/ISBN/article/view/5405

[9] A. Paszke et al., “PyTorch: An imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems,” NeurIPS, no. NeurIPS, pp. 8026–8037, 2019.

[10] I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” 7th Int. Conf. Learn. Represent. ICLR 2019, 2019.

Downloads

Published

2025-07-10

Issue

Vol. 9 No. 3 (2025): Prosiding Seminar Nasional Inovasi Teknologi Tahun 2025

Section

Articles

License

Copyright (c) 2025 John Christofel Wicaksono, Julian Sahertian, Rony Heri Irawan

Creative Commons 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:

  1. Copyright on any article is retained by the author(s).
  2. 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.
  3. 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.
  4. 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.
  5. The article and any associated published material is distributed under the Creative Commons Attribution-ShareAlike 4.0 International License

How to Cite

Klasifikasi Bentuk Wajah Menggunakan Efficientnet-B4. (2025). Prosiding SEMNAS INOTEK (Seminar Nasional Inovasi Teknologi), 9(3), 2119-2129. https://doi.org/10.29407/7dfa6t91