摘要
介绍了一种改进型快速独立分量分析(FastICA)算法与形态学相结合的图像分割方法.该方法把图像的特征分量看作是边缘图像分量与其它背景图像分量的结合,把快速ICA对图像分量的提取,变为对边缘图像分量的提取,得到边缘图像的独立分量,再通过数学形态学的方法对边缘图像进行增强处理,从而实现图像的分割.实验结果表明:与传统的图像分割方法相比,该方法具有良好的图像分割性能,可以清楚地观察到图像轮廓,图像边缘的连通性较好且保留了原图像的很多细节,分割效果较好.
A method for image segmentation based on combination of improved fast independent component analysis and mathematical morphology is proposed.The method regards the independent component of image as the combination of independent component of edge and independent component of background,it translates the extraction of image component into the extraction of edge image component,then get the independent component of the image edge,after enhancing the edge through the mathematical morphology transformation and realize the image segmentation.The result shows that it has better segmentation performance compared with the traditional methods of image segmentation,the image contour can be clearly observed,the segmented image has the good edge connectivity and retains the details of original image.
出处
《微电子学与计算机》
CSCD
北大核心
2015年第2期80-83,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(11174077)
湖南省自然科学基金项目(11JJ3079)
湖南省教育厅基金项目(12C0237)
