摘要
针对传统的高斯混合模型分割方法易受复杂背景影响,且对前景对象的分割效果不理想的问题,提出了一种将高斯混合模型与支持向量机分类器相结合的算法。该算法首先将视频图像由高斯混合模型做初步二值化分割,同时将视频图像用训练后的支持向量机分类器进行像素分类,获取对应的前景块和背景块;然后,将支持向量机得到的分割模板和高斯混合模型分割的结果进行与运算,得到最终分割结果。实验结果表明,该算法显著减少了动态前景对象的分割错误,提高了分割质量。
The conventional Gaussian mixture method suffers from complex backgrounds,which brings about undesirable segmentation result. A novel approach was proposed. It integrated the Gaussian mixture model with a support vector machine classifier. Firstly,this method made preliminary segmentation to get binary image though GMM. Then the pixels in every frame were classified as background and foreground pixels which composed the corresponding background and foreground blocks. At last, the integration of SVM and GMM can be given simply by the intersection operation of segmentation results. Experimental results show this approach significantly decreases the false motion detection and improves segmentation quality of moving objects.
出处
《武汉理工大学学报（信息与管理工程版）》
CAS
2009年第6期857-860,共4页
Journal of Wuhan University of Technology：Information & Management Engineering
基金
国家自然科学基金资助项目(60774051)
