Race Classification from Face Images Using Fast Fourier Transform and Discrete Cosine Transform

Volume 22 , Issue 2 , December 2020 , Pages 149-156

Authors

Hawkar O. Ahmed 1

1 Department of Information Technology, College of Commerce, University of Sulaimani, Sulaimani, Kurdistan Region- Iraq

DOI logo 10.17656/jzs.10816

Keywords

Abstract

Ethnicity identification and recognition is a key biometric technology with a wide range of applications related to homeland security, safety, access control, and automatic annotation. Ethnicity identification from face images is a process of gathering facial features of an individual face image compared to existing face images in the dataset to interpretation his/her ethnic class. In this paper, a propose method in multi-level fusion schema for ethnicity identification by using two global features; fast Fourier transform (FFT) and discrete cosine transform (DCT) on the pre-processed face image of size 128 * 128 in YCbCr color space. A dataset is consisting of 750 face image of three different ethnicities (Kurd 300, Oriental 300 and African 150). The query image feature is compared with a dataset image features using k – nearest neighbor classifier using City block distance for evaluating similarity measurement. The experimental result shows good accuracy and demonstrate the effectiveness of the combined features reached an accuracy rate 96.22% of classification.

Statistics
  • Article view458
  • Downloads0
  • Published at20 December 2020

  • RIS
  • BibTeX
  • EndNote
  • Mendeley
  • APA (7th edition)
  • MLA (9th edition)
  • Chicago
  • Harvard
  • IEEE
  • Vancouver