Eye localization and state analysis of head multi-position in color image
Article
Figures
Metrics
Preview PDF
Reference
Related
Cited by
Materials
Abstract:
To solve the problem of eye localization of human face in multi-angle state of color image,a new method for the location and extraction of the eye image based on color information is proposed,and the method is not influenced by head posture. Firstly,face is detected in color images. Next,the non-skin region in the face area is found based on the difference between skin and eyes/eyebrows. In order to remove the interference points,mathematical calculation is exploited for the non-skin part. Then,eyes are detected with the method of average complexity. The 2-D correlation coefficient is used as a symmetry similarity measure to check the factuality of the 2 detected eyes. Finally,the white and iris of the eyes are precisely located in the eye region,and the eye opening extent can be measured by calculating the eye area. The experimental results demonstrate the proposed method can accurately detect eyes and determine the state of the eye quickly. The method is simple and fast,and has advantage in positioning accuracy and state analysis.