DETECTION SYSTEM OF TEN FINGERPRINT PATTERN USING MATHEMATICAL MORPHOLOGY AND BACKPROPGATION ARTIFICIAL NEURAL NETWORK

  • Wiji Lestari (1*) Universitas Duta Bangsa Surakarta
  • Basiroh Basiroh (2) Universitas Nahdlatul Ulama Alghazali Cilacap, Indonesia
  • Pipin Widyaningsih (3)

  • (*) Corresponding Author
Keywords: fingerprint patterns, mathematical morphology, artificial neural network

Abstract

This research has aim to produce for detection system of human ten fringerprints patterns that according to Dermatoglypic. The fringerprints patterns able to use for advanced analysis to biological and psychological characteristics. This research use back propagation algorithm of neural network to identify of fringerprint patterns. Initial processing is used mathematical morphology method before it is detected. The image is changed to digital image and then it is processed by dilation and erotion for enhancement image. The image that as neuron input of back propagation is changed to gray scale and 8 x 8 of size. Training process use 2000 epochs and patterns [200 2 1]. The output result  are identification of human ten fringerprints patterns. This research produce identification are whorl, arch,  right loop and left loop patterns of fringerprints. The result of research are whorl patterns 51.67%, right loop patterns 23.33%  and left loop 18.33%. The accuration of detection system is 93.33%.

Downloads

Download data is not yet available.
Published
2021-02-25
How to Cite
Lestari, W., Basiroh, B., & Widyaningsih, P. (2021). DETECTION SYSTEM OF TEN FINGERPRINT PATTERN USING MATHEMATICAL MORPHOLOGY AND BACKPROPGATION ARTIFICIAL NEURAL NETWORK. Jurnal Pilar Nusa Mandiri, 17(1), 1-8. https://doi.org/10.33480/pilar.v17i1.1988
Article Metrics

Abstract viewed = 44 times
PDF downloaded = 45 times