Nuttakorn Chalito . Real time face detection using edge and genetic algorithm . Master's Degree(Information Technology). King Mongkut's University of Technology North Bangkok. Central Library. : King Mongkut's University of Technology North Bangkok, 2008.
Real time face detection using edge and genetic algorithm
Abstract:
Face detection has been investigated by many researchers. Nowadays, a lot of
papers state about the skin color model and combining the result with some algorithm
to validate the consequence. They have some far or slightly difference method to
solve the problem. They still abound with the troubles such as lighting condition and
time to process.
This thesis proposes an approach for a face detection system, called real time
face detection using edge and genetic algorithm that aims to improve the quality of
face detection approach.
This algorithm contains three major modules: i) creating a face template ii)
segmentation and iii) template matching. Creating a face template determines average
of pixel value from images. Face segmentation is constructed by using eye
localization and morphology operations for binary image to generate face candidate.
In the final step, each face candidate is verified by Genetic Algorithm (GA).
For evaluation, the experimental testing of detection was conducted using a
database face that belongs to AT&T Laboratories Cambridge, Labeled Faces in the
Wild dataset and images from webcam. The results shown that the accuracy in case
of face folder and non-face folder were 96% and 92%. In both cases, the generation
of GA operator was assigned to 1.