Diffusion and interpolation method for Gradient Vector Flow
Ruzheng Zhao, Zhiheng Zhou, Huiqiang Zhong · 2016
Due to the heavy computational cost and being vulnerable to noise pollution, traditional Gradient Vector Flow(GVF), which is a most popular method in image segmentation, often presents poor performance. A novel generating method of GVF field is presented in this letter. Only a part of image pixels will still use the standard GVF vector diffusion method. More pixels will use the proposed interpolation method. Experimental results show that the proposed method can reduce the computational cost significantly and improve robustness to noise pollution.