Gray Level Face Recognition Using Spatial Features

Md. Fawwad Hussain, Haidong Wang, Kc Santosh

Abstract

Face recognition has always been an active research area with several applications, such as security access control, human-machine interface and gender classification. More often, in real world, grayscale images have been used: video surveillance, for instance. Further, difficulties in face recognition could be due to face poses, orientation, lighting, aging etc. Faces, either in color or grayscale and are having any difficulties (as mentioned earlier) can be learned through edge map and texture, where spatial properties could be learned. Inspired from the fact that face can be considered as line-rich pattern/object, we propose novel face recognition framework that helps learn/recognize via spatial arrangements of edges (and textures as complement). To exploit edge map, we use shape context (SC) and pyramid histogram of orientated gradient (PHOG), and similarly GIST as texture features. Experimental tests (on four different publicly available datasets, such as Caltech, ColorFERET, IndianFaces and ORL) conforms that spatial features are crucial in face representation and recognition.
BibTeX
@inproceedings{hussain2018gray,
  author = {Md. Fawwad Hussain and Haidong Wang and KC Santosh},
  title = {{Gray Level Face Recognition Using Spatial Features}},
  booktitle = {RTIP2R},
  year = {2018},
  doi = {10.1007/978-981-13-9181-1\_20},
}

Loading related content...

Figure