Analisis Pemilihan Optimizer dalam Arsitektur Convolution Neural Network VGG16 dan Inception untuk Sistem Pengenalan Wajah

Ken Ratri Retno Wardani, Hanjaya Suryalim, Ventje J. Lewi Engel, Hans Christian

Abstract


Penggunaan sistem pengenalan wajah semakin meningkat dewasa ini karena itu penting untuk menemukan cara yang optimal dalam meningkatkan akurasi pengenalan wajah. Pengenalan wajah memanfaatkan arsitektur Convolution Neural Network (CNN), tersusun dari lapisan-lapisan konvolusi yang diikuti oleh fully connected layer. Lapisan konvolusi ini yang bertanggungjawab atas proses ekstraksi fitur pada citra yang akan digunakan untuk klasifikasi citra tersebut. Pada penelitian ini diuji dua jenis arsitektur CNN yaitu VGG16 dan Inception untuk mengukur akurasi pengenalan wajah. Faktor lain seperti hyperparameter juga memegang andil tingkat akurasi model. Hyperparameter yang diuji pada penelitian kali ini adalah jenis optimizer dan pengaruh perubahan learning rate pada akurasi. Optimizer bekerja dengan cara mengubah nilai bobot dan bias saat proses backpropagation dengan tujuan menghasilkan nilai error yang minimum. Setiap optimizer memiliki algoritma yang unik. Pengujian menggunakan 2 dataset yaitu Komnet dan Yale, serta melakukan pengujian pengaruh preprocessing MCLAHE terhadap akurasi. Hasil akurasi tertinggi yang dicapai adalah arsitektur Inception dengan optimizer AdaDelta pada dataset Komnet+MCLAHE. Akurasi pada tahap pelatihan mencapai 98%. Rata-rata akurasi setelah model diuji dengan 10-fold cross validation adalah 99.3%.


Keywords


face recognition; CNN; MCLAHE;optimizer; Inception, VGG16

Full Text:

PDF

References


R. I. Bendjilali, B. Mohammed, M. Khaled, and A. Taleb-Ahmed, Illumination-robust face recognition based on deep convolutional neural networks architecture, Indonesian Journal of Electrical Engineering and Computer Science, May 2020.

J. Brownlee, "Why Optimization Is Important in Machine," Machine Learning Mastery, [Online] https://machinelearningmastery.com/. [Accessed: February 12, 2022].

M. Yaqub, J. Feng, S. Zia, A. Arshid, K. Jia, U. R. Zaka, and A. Mehmood, "State-of-the-Art CNN Optimizer for Brain Tumor Segmentation in Magnetic Resonance Images," in Brain Sciences, published July 3, 2020.

S. S. Mohammed, W. A. Mohamed, A.T. Khalil, and A.S. Mora, "Deep Learning Face Detection and Recognition," in International journal of electronics and telecommunications, June 2019.

K. B. Pranav and J. Manikandan, "Design and Evaluation of a Real-Time Face Recognition System using Convolutional Neural Networks," Article in Procedia Computer Science 171, 1651–1659, 2020.

M. Coskun, A. Ucar, O. Yildrim, and Y. Demir, "Face Recognition Based on Convolutional Neural Network," in Conference Paper, November 2021.

Z. Xie, J. Li, and H. Shi, "A Face Recognition Method Based on CNN," in Journal of Physics: Conference Series, 2019, doi: 10.1088/1742-6596/1395/1/012006.

M. K. Hasan, M. S. Ahsan, A. A. Mamun, S. H. S. Newaz, and G. M. Lee, "Human Face Detection Techniques: A Comprehensive Review and Future Research Directions," Electronics 2021, 10, 2354, [Online] https://doi.org/10.3390/electronics10192354.

L. Patil and V.D. Mytri, "Face Recognition with CNN and Inception Deep Learning Models," International Journal of Recent Technology and Engineering (IJRTE), Volume-8 Issue-3, September 2019.

H. E. Khiyari and H. Wechsler, "Face Recognition across Time Lapse Using Convolutional Neural Networks," in Journal of Information Security, Vol.7 No.3, 2016.

K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition," in The 3rd International Conference on Learning Representations ICLR, 2015, [Online] https://arxiv.org/abs/1409.1556. [Accessed: January 20, 2022].

A. Rosebrock, Deep Learning for Computer Vision with Python, PyImageSearch. 2017.

M. Davidson, "Review on Gradient Descent Algorithms in Machine Learning," in Journal for Innovative Development in Pharmaceutical and Technical Science, Volume-2, Issue-10, October 2019, [Online] https://ssrn.com/abstract=3810947.

S. Ruder, "An overview of gradient descent optimization algorithms", [Online] https://www.semanticscholar.org/reader/769ef3d5021cd71c37d2c403f231a53d1accf786. [Accessed: February 1, 2022, 16:28:00].

Vidushi et al., "Assessment of Optimizers impact on Image Recognition with Convolutional Neural Network to Adversarial Datasets," in Journal of Physics: Conference Series, 2021, doi:10.1088/1742-6596/1998/1/012008.

K. Naik, "Understanding All Optimizers In Deep Learning," [Online] https://krishnaik.in/2022/03/28/understanding-all-optimizers-in-deep-learning. [Accessed: February 7, 2022, 23:41:00].

Y. Ding, "The Impact of Learning Rate Decay and Periodical Learning Rate Restart on Artificial Neural Network," in Proceedings of the 2021 2nd International Conference on Artificial Intelligence in Electronics Engineering, pp. 6–14, 2012, [Online] https://doi.org/10.1145/3460268.3460270.

P. Huilgol, "Accuracy vs F1-Score," [Online] https://medium.com/analytics-vidhya/. [Accessed: 12 February 2022, 22:39:00].




DOI: http://dx.doi.org/10.26418/jp.v9i2.60432

Refbacks

  • There are currently no refbacks.