Title :
A Novel Variational Segmentation Method for High-Noise Image
Author :
Luo, Zhihong ; Feng, Guocan ; Yang, Guan
Author_Institution :
Sch. of Inf. Sci. & Technol., Sun Yat-sen Univ., Guangzhou, China
Abstract :
A novel variational method coupling denoising and segmentation is proposed. First, the high-noise images are denoised by the proposed denoising model, then they are segmented by the new segmentation model. And through multiplication operator splitting (MOS) numerical algorithm, the effective and stable results are obtained. Experimental results clearly demonstrate that the proposed method coupling denoising and segmentation has effectiveness and it has an advantage over CV model in high-noise image segmentation.
Keywords :
image denoising; image segmentation; mathematical operators; variational techniques; high-noise image segmentation; image denoising; multiplication operator splitting; numerical algorithm; variational segmentation method; Image denoising; Image segmentation; Information science; Mathematics; Noise reduction; Speckle; Sun;
Conference_Titel :
Pattern Recognition, 2009. CCPR 2009. Chinese Conference on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4244-4199-0
DOI :
10.1109/CCPR.2009.5344063