Estimasi Gender Berbasis Sidik Jari dengan Wavelet dan Support Vector Machine

Penulis

  • Sri Suwarno Universitas Kristen Duta Wacana, Yogyakarta
  • Aditya Wikan Mahastama Universitas Kristen Duta Wacana, Yogyakarta

DOI:

https://doi.org/10.25126/jtiik.1077972

Kata Kunci:

gender, sidik jari, Haar wavelet, SVM

Abstrak

Estimasi gender berbasis sidik jari sering diperlukan untuk identifikasi jenazah tanpa identitas, sebelum dipastikan dengan tes DNA.  Untuk kepentingan tersebut sidik jari jenazah diambil secara digital dan selanjutnya diidentifikasi. Kesulitan yang dihadapi dalam memproses sidik jari secara digital adalah menentukan fitur yang handal dan tidak dipengaruhi oleh kwalitas citra dan masalah translasi dan rotasi.  Selain itu untuk mendapatkan akurasi yang tinggi diperlukan sejumlah preprocessing pada dataset. Penelitian ini bertujuan untuk mengestimasi gender berbasis sidik jari dengan fitur yang dibangkitkan dari transformasi wavelet. Fitur diambil dari nilai Energy yang dihasilkan melalui transformasi Haar wavelet sebanyak enam level. Selanjutnya fitur tersebut dipakai sebagai data latih bagi Support Vector Machines (SVM) untuk diestimasi. Penelitian ini menggunakan dataset dari NIST (National Institute of Standart and Technology) sebanyak 1000 sampel terdiri dari 500 sidik jari pria dan 500 sidik jari wanita.  Berdasarkan hasil percobaan yang dilakukan, metode ini menghasilkan akurasi sampai 70.3% dengan tingkat TPR (True Positive Rate) sebesar 80.6% untuk sidik jari wanita dan 60.0% untuk sidik jari pria. Metode ini menunjukkan waktu komputasi yang cepat karena tidak memerlukan preprocessing, komputasinya sederhana dan dengan jumlah sampel data yang tidak banyak.

 

Abstract

 

Gender estimation based on fingerprints is often required to identify unidentified corpses before being confirmed by DNA testing. For this purpose, fingerprints of the body are taken digitally and subsequently identified. The difficulty in digital fingerprint examination is determining reliable features unaffected by translation or rotation. In addition, some preprocessing is required on the dataset to obtain high accuracy. This study aims to estimate gender based on fingerprints using wavelet transform and Support Vector Machines (SVM). The features are the  Energy values generated by the Haar wavelet transform of six-level. The features are then used as training data for the SVM to be classified. This study used datasets from NIST (National Institute of Standards and Technology), as many as 1000 samples consisting of 500 male fingerprints and 500 female fingerprints. Based on the experiments' results, this method produces an accuracy of up to 70.3% with a TPR (True Positive Rate) of 80.6% for female fingerprints and 60.0% for male fingerprints. This method shows a fast computational time because it does not require preprocessing; the computation is simple and with a small amount of sample data.

Downloads

Download data is not yet available.

Referensi

GNANASIVAM, P. and VIJAYARAJAN, R. 2019. Gender classification from Fingerprint Ridge Count and fingertip size using optimal score assignment, Complex & Intelligent Systems, 5(3), pp. 343–352. doi:10.1007/s40747-019-0099-y.

HSIAO, C.-T. et al. 2022. Application of convolutional neural network for fingerprint-based prediction of gender, finger position, and height, Entropy, 24(4), p. 475. doi:10.3390/e24040475.

ILOANUSI, O.N. and EJIOGU, U.C. 2020. Gender classification from fused multi-fingerprint types, Information Security Journal: A Global Perspective, 29(5), pp. 209–219. doi:10.1080/19393555.2020.1741742.

JAKALALA, G. and SUDHA, L.R. 2021. Gender Classification Based on Fingerprint Analysis, Turkish Journal of Computer and Mathematics Education (TURCOMAT), 12(10), pp. 1249–1256. doi:https://doi.org/10.17762/turcomat.v12i10.4318.

JALALI, S., Boostani, R. and Mohammadi, M. 2021. Efficient fingerprint features for gender recognition, Multidimensional Systems and Signal Processing, 33(1), pp. 81–97. doi:10.1007/s11045-021-00789-6.

LEE, L.C. et al. 2020. Forensic gender discrimination in Malaysian population using machine learning methods, 2020 International Conference on Innovation and Intelligence for Informatics, Computing and Technologies (3ICT)[Preprint]. doi:10.1109/3ict51146.2020.9311988.

RIM, B., KIM, J. and HONG, M. 2020. Gender classification from fingerprint-images using Deep Learning Approach, Proceedings of the International Conference on Research in Adaptive and Convergent Systems [Preprint]. doi:10.1145/3400286.3418237.

KRUTHI, R,. PATIL, A. and GORNALE, S. 2019. Fusion of local binary pattern and local phase quantization features set for gender classification using fingerprints, International Journal of Computer Sciences and Engineering, 7(1), pp. 22–29. doi:10.26438/ijcse/v7i1.2229.

SUWARNO, S dan KURNIAWAN, E. 2023. ‘Multi-Level Pooling Model for Fingerprint-Based Gender Classification, MATRIK : Jurnal Manajemen, Teknik Informatika Dan Rekayasa Komputer, 22(2), pp. 195-206. doi: 10.30812/matrik.v22i2.2551.

TERHORST, P. et al. 2018. What can a single Minutia tell about gender?, 2018 International Workshop on Biometrics and Forensics (IWBF) [Preprint]. doi:10.1109/iwbf.2018.8401554.

WALKER, J.S. 2008. A Primer on wavelets and their scientific applications. Boca Raton, FL: Chapman & Hall/CRC.

Diterbitkan

30-12-2023

Cara Mengutip

Estimasi Gender Berbasis Sidik Jari dengan Wavelet dan Support Vector Machine. (2023). Jurnal Teknologi Informasi Dan Ilmu Komputer, 10(7), 1431-1436. https://doi.org/10.25126/jtiik.1077972