کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
848810 | 909252 | 2014 | 5 صفحه PDF | دانلود رایگان |
In this paper, a new color balloon snake model is introduced and used for face segmentation in color images. It is an extension of existing balloon snake models. Based on a coarse detection of facial features, the method combines a skin-tone distribution model and a boundary diffusion model to search for the facial boundary. The skin distribution is a single Gaussian, which is proposed to extract the skin-tone region in the RGB space. The diffusion model, which is invented to diffuse the facial boundary, is a one-dimensional Gauss revolution surface. The parameters are evaluated based on an AdaBoost face detection method. The color snakes are weighted by the distributions, and the external forces evolve dynamically to reach the boundary, which depends on the balance between the internal and external forces. Experiments were conducted, and the results show that the model provides desired segmentation outcomes. It is robust against complex backgrounds and lighting pollution.
Journal: Optik - International Journal for Light and Electron Optics - Volume 125, Issue 11, June 2014, Pages 2538–2542