时空联合的视频对象分割，时域采用多帧帧差，形态学处理和高斯聚类，空域采用小波变换和分水岭变换。时域分割的模板与空域分水岭变换的结果融合得到最终的分割结果。-joint space-time video segmentation, multi-domain when displayed poor handling and morphology Gaussian clustering, airspace wavelet transform and watershed transformation.

基于小波的视频分割，经过时域和空域来综合获得清晰对象轮廓，再进行行扫面添白-Wavelet-based video segmentation, through time-domain and the airspace to clear and comprehensive access to the object contour, and then line scanning Tim White