1. Infrared Face Recognition Based on Histogram and K-Nearest Neighbor Classification.
- Author
-
Wang, Shangfei and Liu, Zhilei
- Abstract
Infrared facial images record the temperature-field distribution of facial vein embranchment, which can be regarded as gray features of images. This paper proposes an infrared face recognition algorithm using histogram analysis and K-Nearest Neighbor Classification. Firstly, the irregular facial region of an infrared image is segmented by using the flood-fill algorithm. Secondly, the histogram of this irregular facial region is calculated as the feature of the image. Thirdly, K-Nearest Neighbor is used as a classifier, in which Histogram Matching method and Histogram Intersection method are adopted respectively. Experiments on Equinox Facial Database showed the effectiveness of our approach, which are robust to facial expressions and environment illuminations. [ABSTRACT FROM AUTHOR]
- Published
- 2010
- Full Text
- View/download PDF