加权三维Otsu方法在图像分割中的应用*

从信息散度角度出发，提出一种新的三维Otsu加权方法，对分布不均匀的图像做细致分割。由于三维Otsu中三个维度在分割中所起的作用不同，可以通过调整不同维度的权因子，使不同的图像都能得到很好的分割效果。最后利用加权的特点，将三维Otsu分解为三个一维Otsu，在加快运算速度，降低时间复杂度的同时得到了比较好的分割效果，使得加权方法更具有实际效用。

Application of weighting 3D-Otsu method in image segmentation

From the point of view of information divergence, a new weighting method of 3D-Otsu is proposed in this paper, witch which a precise segmentation result for inhomogeneous image can be obtained. Since the three dimensions in 3D-Otsu method play different roles in segmentation of images, better results of segmentation can be obtained by changing the weighting factors of the three dimensions in 3D-Otsu method. According to the characteristic of weighting method, the 3D-Otsu can be decomposed into three one-dimensions. So, this method can speed up the calculation and lower the computational complexity and make the weighting method more practical. With this method, an optimized segmentation result is obtained in this paper.