Click for new scientific resources and news about Corona[COVID-19]

Paper Information

Journal:   AMIRKABIR   FALL 2007-WINTER 2008 , Volume 18 , Number 67-A (TOPICS IN: GROUPS OF ENGINEERING, ELECTRONIC, BIOMEDICAL, COMPUTER); Page(s) 83 To 92.
 
Paper: 

FACE DETECTION BY GABOR TRANSFORM

 
 
Author(s):  NOROUZI M.R., VAFADOUST M.
 
* 
 
Abstract: 

A single reference image is used for detecting the face area in gray scale images. The algorithm uses Gabor transform as a powerful tool for feature extraction. Zero-crossing of Gabor transform is used for constructing the binary feature vectors and Hamming distance is used for measuring the similarity between feature vectors. Our developed method uses a pyramidal structure for finding the face images with different sizes. The algorithm has considerable robustness to the cluttered backgrounds and non-uniform illumination. More than 11,000 face images have been tested. The results show a very high detection rate for non occluded face images with low facial expression. The algorithm can also be used for correcting the head rotation angle in input images. The algorithm can also be used for correcting the head rotation angle in input images. The proposed method is suitable for the cases that contain one face image.

 
Keyword(s): FACE DETECTION, GABOR WAVELET, GABOR TRANSFORM, HAMMING DISTANCE
 
References: 
  • ندارد
 
  Yearly Visit 67
 
Latest on Blog
Enter SID Blog