Title :
Skin Color Segmentation Based on Improved 2D Otsu and YCgCr
Author :
Zhao, Minghua ; Zhao, Yonggang
Author_Institution :
Fac. of Comput. Sci. & Eng., Xi´´an Univ. of Technol., Xi´´an, China
Abstract :
A new algorithm for skin color segmentation is proposed in this paper. Firstly, the skin color sample images are light-compensated and transferred from RGB to YCgCr color space. Secondly, the Gaussian skin color model is established from 179221 skin pixels. Thirdly, the skin like similarity degree is computed and the skin color similarity image is obtained. Finally, an improved 2D Otsu method is used into skin color segmentation. Experimental results show that the new skin color segmentation method is superior to the traditional methods.
Keywords :
Gaussian processes; image colour analysis; image segmentation; skin; 2D Otsu method; Gaussian skin color model; RGB; YCgCr color space; light-compensated; skin color sample images; skin color segmentation; skin color similarity image; skin pixels; Computational modeling; Equations; Image color analysis; Image segmentation; Mathematical model; Pixel; Skin; 2D Otsu method; color space; skin color; skin color segmentation;
Conference_Titel :
Electrical and Control Engineering (ICECE), 2010 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-6880-5
DOI :
10.1109/iCECE.2010.480
