H. C. VijayLakshmi and S. Patilkulkarni DOI: 10.7763/IJCTE.2013.V5.646 56 Face Detection in Skin-Toned Images Using Edge Detection and Feature Extraction Using R and G Channels through Wavelet Approximation Abstract—Face detection and localization in complex skin toned background is a highly challenging problem. In this paper, use of combination of color spaces and edge detection in red and green channels is proposed for segmenting out the skin-tone regions. Wavelet approximations are used for the extraction of prominent features of face. Experimental results are shown to yield the improved false acceptance rates (FAR) over the algorithms that either use grey scale image for segmentation and the algorithms that do not use any edge detection. Index Terms—Face detection, localization, edge, wavelet, r and g channel. I. INTRODUCTION Face detection and localization is the task of checking whether the given input image contains any human face, and if so, returning the location of the human face in the image. Face detection is difficult mainly due to a large component of non-rigidity and textural differences among faces. The great challenge for the face detection problem is the large number of factors that govern the problem space [1], [2]. The long list of these factors include the pose, orientation, facial expressions, facial sizes found in the image, luminance conditions, occlusion, structural components, gender, ethnicity of the subject, the scene and complexity of image‟s background. The scene in which the face is placed ranges from a simple uniform background to highly complex backgrounds. In the latter case it is obviously more difficult to detect a face. Faces appear totally different under different lighting conditions. A thorough survey of face detection research work is available in [1],[2]. In terms of applications, face detection and good localization is an important preprocessing step in online face recognition. II. CHALLENGES For the problem of face detection involving colour images having complex scenes, the use of skin pixel properties for segmentation reduces the search space to a greater extent [3]. Most of the detection algorithms have considered images having non-skin tone background, people wearing non-skin Manuscript received June 25, 2012; revised July 30, 2012. H. C. Vijaylakshmi is with Sri Jayachamaranendra College of Engineering, Mysore. Karnataka, India (e-mail [email protected]). Dr. Sudarshan PatilKulkarni is with JSS Research Foundation, SJCE , Mysore, Karnataka, India (e-mail [email protected]). tone dresses etc. If images contain skin tone background, then the entire region is identified as skin region “Fig. 1a” and “Fig. 1b”. In order to locate faces present in the segmented regions calls for additional face localization process. While segmenting faces of people wearing skin-tone dresses using skin pixel segmentation preprocessing technique, the entire image of the person with skin-tone dress is detected as the face region and hence requires a further face localization step. Besides, overlapping face regions also add additional constraints while segmenting the faces. Due to variation in illumination, skin regions may not be identified properly as skin during skin segmentation. Locating faces in these circumstances is more complex as opposed to localizing faces with uniform, non skin-tone background. Inspite of using combination of different colour spaces during segmentation, it is tedious to demarcate region boundaries between skin and pseudo skin regions and also eliminate these regions from searching process. The use of colour space alone sometimes fails to segment the boundary regions of the image. In order to overcome this problem combination of colour spaces for efficient skin segmentation followed by Canny and Prewitt edge detection to demarcate the region boundary is used for input image segmentation [4]. Fig. 1 a. Input image with skin tone background Fig. 1 b. Segmented image using HSI and YCbCr combination. III. SEGMENTATION AND FEATURE EXTRACTION Segmentation and feature extraction are the two important pre-processing steps that play a vital role in face detection International Journal of Computer Theory and Engineering, Vol. 5, No. 1, February 2013
4
Embed
Face Detection in Skin-Toned Images Using Edge …ijcte.org/papers/646-W00097.pdf · Detection and Feature Extraction Using R and G Channels ... face detection, skin segmentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
H. C. VijayLakshmi and S. Patilkulkarni
DOI: 10.7763/IJCTE.2013.V5.646 56
Face Detection in Skin-Toned Images Using Edge
Detection and Feature Extraction Using R and G Channels
through Wavelet Approximation
Abstract—Face detection and localization in complex skin
toned background is a highly challenging problem. In this
paper, use of combination of color spaces and edge detection in
red and green channels is proposed for segmenting out the
skin-tone regions. Wavelet approximations are used for the
extraction of prominent features of face. Experimental results
are shown to yield the improved false acceptance rates (FAR)
over the algorithms that either use grey scale image for
segmentation and the algorithms that do not use any edge
detection.
Index Terms—Face detection, localization, edge, wavelet, r
and g channel.
I. INTRODUCTION
Face detection and localization is the task of checking
whether the given input image contains any human face, and
if so, returning the location of the human face in the image.
Face detection is difficult mainly due to a large component of
non-rigidity and textural differences among faces. The great
challenge for the face detection problem is the large number
of factors that govern the problem space [1], [2]. The long list
of these factors include the pose, orientation, facial
expressions, facial sizes found in the image, luminance