Finding the estimated position of facial features on the human face using intensity computation

Most human face recognition approaches start with a detection of facial features (facial feature extraction). To extract the right facial features, a process for finding their exact position is needed. The aim of this paper is to find the estimated position of facial features using intensity computation and to explain pre-processing in facial feature extraction. For the computerized images used in this research, pixels were manipulated in intensity only, without changing their original position. The system performance was tested on 40 test data (20 good quality images and 20 interference images), locating eyes and lips line only. Using good quality images, the system can show all facial features expected and from interference images, the system only gives 90% of accuracy. Intensity computation is one possible method to find the estimated positions of facial features. The corner selection and capturing area processes give accurate results for locating the eyes and lips line