基于扩散方程的强度不均匀图像分割

Intensity inhomogeneity image segmentation based on diffusion equation

图像的强度不均匀性往往会导致目标边缘模糊,以致难以获得一个令人满意的图像分割结果.为获得较精确的分割结果,该文提出了一种基于扩散方程的强度不均匀图像分割模型.该模型无须设置初始轮廓,直接对图像进行类Laplace算子预处理构造初始的冷、热源,根据热传导的思想对强度严重不均匀的图像进行有效的分割.通过与多种水平集图像分割模型进行实验对比,发现文中的模型在强度不均匀图像的分割精度上基本优于其他模型;运行时间也具有明显优势.实验表明,文中方法针对强度不均匀图像的分割具有高效率、高精度、高鲁棒性的优点.

The intensity inhomogeneity of the image often leads to the fuzzy edge of the target,which makes it difficult to obtain a satisfactory image segmentation result.In order to obtain more accurate segmentation results,a intensity inhomogeneity image segmentation model based on diffusion equation is proposed in this paper.This model does not need to set the initial contour,and directly constructs the initial cold and heat source by laplace-like pre-processing of the image.According to the idea of heat conduction,the image with severely intensity inhomogeneity is effectively segmented.Compared with many level set image segmentation models,it is found that the proposed model is superior to other models in the segmentation accuracy of images with intensity inhomogeneity.Uptime also has obvious advantages.Experiments show that the proposed method has the advantages of high efficiency,high accuracy and high robustness for image segmentation with intensity inhomogeneity.