Skin Color Based Segmentation for Multiple Face Detection

Title: Skin Color Based Segmentation for Multiple Face Detection
Publisher: Guru Nanak Publications
ISSN: 2278-0947
Series: Volume 4 Issue 2
Authors: T. Archana , Dr. T. Venu Gopal and M. Praneeth Kumar


Nowadays, human face is being used as an index for identification. In present days automatic face detection has become an interesting research field. Face detection in an image is “to separate human frontal faces from the background and determine their location in it regardless of their position, scale, in plane rotation, pose, illumination etc”. Automating an application process related to image processing needs image analyzing methods. This paper presents a novel face detection approach that uses skin color based segmentation and morphological processing. To perform face detection , the algorithm uses extraction of color planes, thresholding, erosion and dilation morphological operations , filtering (for avoiding false detection). Next, particle analysis is performed to identify or locate only the face skin area in the given image and not the other parts of the body. The color planes are extracted using vision module the RGB represented color space is converted into a suitable color model such as YCbCr. The method presented can be applied to detect or identify single as well as multiple human faces in the given input image. Face recognition is the process of identifying a given input image is available in the stored image database. Face detection is one of the essential preprocessing step for any face recognition approach. Template based face recognition is one of face recognition approach. In this paper the performance of template based face recognition without and with face detection using color based segmentation is measured to evaluate the efficiency and the effectiveness of face detection. Experimental results conducted on FERET database show that the algorithm is efficient in detecting the human faces with an accuracy of 100% for most of the cases and it is also observed that the performance of face recognition is also improved.


Face detection, color based segmentation, extracting color planes, thresholding, morphological operations, erosion and dilation, particle analysis, face recognition.

Download Full Text

(For complimentary copy, please contact