基于改进的JSEG算法的图像分割

针对JSEG算法在图像分割中出现的明显过分割现象,提出一种基于边缘信息的JSEG1]改进方法。该方法首先将图像的颜色空间转换为LUV颜色空间,用PGF（Peer Group Filtering）2]算法对图像进行平滑去噪,用分裂算法确定图像的类数,用GLA（Generalized Lloyd Algorithm）3]算法完成量化,生成＂类图＂。然后计算每个像素的J值,并利用Canny算子检测的边缘信息,对J值进行修正,计算每个像素的局部相似程度,并在不同的尺寸下构建J图像,这样就能反映出最有可能的边界位置。最后在J图像上进行种子区域增长,直到获得最终的分割结果。实验结果表明该方法可以有效地改善JSEG算法在图像分割中存在的过分割现象。

Image Segmentation Based on Improved JSEG Algorithm

This paper proposed an improved JSEG algorithm based on edge information in terms of the over segmentation of image by JSEG algorithm.In our method,we switch the space of color of image to LUV space,and then make ″Class Graph″ by qualification through GLA algorithm.Secondly,count J of every pixel and modify the value of J using edge information detected by Canny algorithm.Counting the partial similarity of image pixel and making J graph in different scales,this paper find out the most possible edge position.At last,we can get the segmentation result through seed region growing of J graph.