摘要
提出了一种基于图象颜色和细节信息的多尺度视频对象提取算法。在低尺度下根据颜色信息进行初始粗分割;随着尺度的增加,利用对象的细节信息对分割结果进行细调整,直到达到需要的精度;最后,利用形态学的有关算法对分割区域进行填充、平滑。实验证明,算法的提取速度快、定位准确,可以用于视频序列中关键帧(I帧)的对象提取。
A novel muhiscale video object extraction scheme based on color and detail information is proposed in this paper， First，a color clustering method is performed to segment images at a coarse scale，Second，the high frequency wavelet coefficients and their statistics are used to perform classification of image blocks at fine scales.Post-processing will refine the boundaries of foreground and eventually extract the objects.The major contribution of this paper is to exploit color information to relieve the influence of complex background and combine the detail information to extract the objects.Experiments show that the new algorithm provides high accuracy and the computation time is quite short。
出处
《计算机工程与应用》
CSCD
北大核心
2005年第23期63-66,共4页
Computer Engineering and Applications
基金
国家部委基金项目资助
关键词
对象提取
多尺度
颜色空间
小波变换
object extraction,muhiscale,color space,wavelet transform
