Article details

Title: Decision Fusion for Face Detection in Color Images
Author(s): Romana Oancea   Cristian Molder   Marius Cîrmaci         

Abstract: Face recognition in color images is important in various applications, such as smartcards, information security, law enforcement, entertainment, and so on. The recognition is divided into two main stages: (1) face detection in images, followed by (2) face recognition itself. Concerning the face detection step, the vast majority of studies use only one algorithm applied using a single color space, with limited performances. In this article, we propose the simultaneous use of several well-known color spaces (such as nRGB, YCbCr, HSV or YES) for constructing skin/non-skin models, followed by the use of decision fusion in every color space considered individually in order to enhance the detection performance. The results are presented considering a database of more than 600 color images of 480×720 pixels, with human subjects in various conditions and backgrounds.

Keywords: image processing, face detection, biometry.

References:

[1] V. VEZHNEVETS, A. ANDREEVA – A Comparative Assessment of Pixel-Based Skin Detection Methods, Technical Report, Moscow State University, Moscow, Russia, 2005
[2] L.I. KUNCHEVA – Switching between Selection and Fusion in Combining Classifiers: An Experiment, IEEE Transactions on Systems, Man, and Cybernetics. Part B: Cybernetics, Vol. 32, No. 2, pp. 146-156, Apr. 2002
[3] M. TABB, N. AHUJA – Multiscale Image Segmentation by Integrated Edge and Region Detection, IEEE Transactions on Image Processing, Vol. 6, No. 5, pp. 642-655, May 1997
[4] A. KHAPARDE, S. REDDY, S. RAVIPUDI – Face Detection Using Color Based Segmentation and Morphological Processing: A Case Study, Proc. of the International Symposium on Computer Engineering and Technology, ISCET 2010, pp. 147-151, Mandi Gobindgarh, Punjab, India, Mar. 19-20, 2010