基于三维直方图的Fisher评价函数图像分割方法

3D histogram based Fisher criterion function for image segmentation

提出了一种基于三维直方图的Fisher评价函数图像分割方法,该方法利用了图像像素点的灰度信息和邻域的均值信息及中值信息,给出了三维阈值分割方法。实验结果表明,相比较基于二维直方图的Fisher线性判别函数图像分割方法,在图像受高斯噪声干扰的情况下,该方法具有更好的分割效果。

In this paper, a Fisher criterion function based on 3D histogram is presented for image segmentation. The 3D threshold segmentation method is given by use of the gray-level trait of the pixels, and the average and median value of the neighboring regions. Compared with the Fisher linear criterion function based on 2D histogram, the new method is better in image segmentation for images with Gaussian noise.