Mobile devices application : alive face detection using skin color segmentation and gabor filter

Abstract EN

Interest in mobile phone applications is increasing.

The circulation of mobile phones with digital cameras led to a number of applications being generated.

Face detection techniques are the first process in many applications and it is not an easy process.

The proposed application method locates a human face by obtaining a capture frame from a camera in live video.

To start processing this image, the application searches for important features and then uses these features to decide the face location.

A popular technique based on skin color is a robust and faster approach than many other techniques for face detection because color has more information processing than other facial feature extraction.

This paper recommends a method for mobile device application to detect the human face by using a skin color model based on YCbCr chrominance space.

Only the frontal face use, segment all skin that detects include all object, and then using an erosions function to delete additional objects that are not skin.

Subsequently, a Gabor filter is used to classify skin color (face or no face).

Our method achieves 96% accuracy and can locate a single face closer to the camera.

It can detect a face in different light conditions in addition to those based on light control in an intricate image background.