Face Detection Essay

1743 Words4 Pages

Face Detection and Facial Feature Extraction Based on a fusion of Knowledge Based Method and Morphological Image Processing

Detecting human faces and extracting the facial features from an image is a challenging process. It is very difficult to locate faces in an image accurately. There are several variables that affect the performance of the detection methods, such as wearing glasses, skin color, gender, facial hair, and facial expressions etc. We propose an efficient method for locating a face region and extracting facial features based on the characteristics of eye regions. The method proposed in this paper is based on the assumption that the frontal face image is available. The face regions are detected from a pair of possible eye candidates. Then, the facial features are extracted from the detected face regions. Our method for detecting and extracting the facial features is divided into two stages. At first, the eye pairs are detected by testing all possible eye regions in an image. After detecting a pair of eye candidates, the distance between the eyes is used to find a possible face candidate. Next, the face is divided into different regions and facial features are extracted from these regions. The extracted features consist of the eye corners, the iris, the nostrils, and the mouth corners.
I. INTRODUCTION
Human face detection and face image analysis have become one of the most important research topics in the world of pattern recognition and computer vision. The eye is the most important feature in a human face. The facial feature detection techniques aim to extract specific features such as, pupil, corners of the eyes, nostrils, lip corners, etc. Major applications of face detection consist of topics such as, face recog...

... middle of paper ...

...detecting the dark regions in the image. Based on the detected pair of eye candidates, possible facial regions are located by means of the geometric relationship between various features. Detection of eyes, mouth and nose are done by estimating the probable region for each feature. Geometrical interpretation of location of facial features, used in the algorithms is described with pictorial representation. It is observed that, with the use of facial geometry, the accuracy of features (eyes, nose and mouth) detection is greatly improved. The proposed method for feature extraction is also found to be accurate in detecting all kinds of frontal images. In conclusion, this method can achieve a high performance in detecting human faces and extracting facial features.
The system was tested on 25 people, and was successful with 20 people, resulting in 80% accuracy (Fig.5).

Open Document