基于时间轴小波变换的运动对象分割算法研究

对时间轴一维小波变换的视频运动对象分割算法进行了研究。将视频序列进行时间轴一维小波变换，利用变换后的高频帧信息提取出初步的运动掩模图像；进行数学形态学后处理，以消除各种噪声的影响和对象的不连通性，得到理想的运动掩模图像。实验结果表明该方法分割效果较好，分割出的运动对象可直接应用于基于对象的视频编码，且算法耗时较少。

Research on motion object segmentation algorithm based on temporal decomposition by wavelet transform

Motion object segmentation algorithm based on temporal decomposition by wavelet transform is explored. The original motion mask image can be extracted from the high frequency frames which resulting from temporal decomposition of the original video sequence by wavelet transform. Then a morphological filter is applied to refine the motion mask which can eliminate the noise and fill the interior hole. Experimental results show that the proposed algorithm has not only better segmentation performance but also low complexity; furthermore, the motion object can be used in the video object coding directly.