Klasifikasi Ras Berdasarkan Citra Wajah Menggunakan Convolutional Neural Network
DOI:
https://doi.org/10.34818/eoe.v7i1.11925Abstrak
Abstrak Salah satu fungsi penting identifikasi ras dalam kehidupan sehari-hari adalah untuk keperluan kontrol kesehatan yang dilakukan oleh pemerintah. Namun, tidak mudah untuk menentukan ras dari masing-masing orang secara langsung, terlebih apabila jumlah masyarakat yang banyak sehingga juga akan memakan waktu lama. Salah satu cara yang dapat dilakukan untuk menyelesaikan masalah tersebut adalah dengan menggunakan sistem otomatis. Sistem otomatis disini adalah berupa perangkat lunak yang dapat digunakan untuk mengidentifikasi ras berdasarkan citra wajah seseorang. Dalam penelitian ini metode yang digunakan adalah Convolutional Neural Network (CNN). Melalui penelitian ini, didapatkan sebuah struktur CNN yang optimal untuk mengklasifikasikan ras. Struktur CNN tersebut memberikan akurasi training sebesar 91.63% setelah dilatih dengan sekitar 5400 citra wajah dan akurasi testing sebesar 89.739% untuk 600 citra wajah. Selain itu, ketika diuji dengan dataset citra baru, CNN tersebut menghasilkan akurasi klasifikasi sebesar 76.67% untuk 30 citra wajah. Kata kunci : ras, klasifikasi, citra wajah, convolutional neural network, CNN Abstract One of the main purpose of race identification in daily life is for the government health control program. But, it is not easy to determine the race of each person directly, it will take a long time, especially because high number of people in the society. One of the technique that could be implemented to solve this problem is by utilizing an automated software system that could identify the person’s race based on their face image. For this research, the Convolutional Neural Network (CNN) is used for the software. The result of this research is an optimal structure of CNN for race classification. The best CNN structure, give 91.63% as the training accuracy value after trained using around 5400 facial images and 89.739% as the testing accuracy value on 600 facial images. Also, when the CNN is tested to classify a new face dataset, it gives 76.67% classification accuracy on 30 facial images. Keywords: race, classification, face image, convolutional neural network, CNNUnduhan
Referensi
[1] L. Agassiz. The Diversity of Origin of the Human Races. Boston Christian Examiner, 1850.
[2] A. Ahmed, W. Xu, K. Yu, and Y. Gong. Training hierarchical feed-forward visual recognition models using transfer learning from pseudo-tasks. Computer Vision – ECCV 2008: 10th European Conference on Computer Vision, 3:69–82, 2008.
[3] L. M. Alcoff. Visible Identities : Race, Gender and the Self. Oxford University Press, 2006.
[4] M. Demirkus, S. Guler, and K. Garg. Automated person categorization for video surveillance using soft biometrics. Proceedings of SPIE - The International Society for Optical Engineering, 2010.
[5] S. Fu, H. He, and Z.-G. Hou. Learning race from face : A survey. IEEE, 2014.
[6] S. Haslanger. Gender and race: (what) are they? (what) do we want them to be? Massachusetts Institute of Technology, 2000.
[7] J. T. Kubota, M. R. Banaji, and E. A. Phelps. The neuroscience of race. National Center for Biotechnology Information, 2012.
[8] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffiner. Gradient-based learning applied to document recognition. IEEE, 1998.
[9] H. Lin, H. Lu, and L. Zhang. A new automatic recognition system of gender, age and ethnicity. 2006 6th World Congress on Intelligent Control and Automation, 2:9988–9991, 2006.
[10] V. M. Mays, N. A. Ponce, D. L. Washington, and S. Cochran. Classification of race and ethnicity: Implications for public health. Annu Rev Public Health, 24:83–110, 2003.
[11] B. S. Ryan. Alleviating own-race bias in cross-racial identifications. Washington University Jurispresidence Review, 8:124–128, 2015.



