一种从视频压缩码流中精确提取运动对象的快速算法

A Fast and Accurate Moving Object Extraction Method from Compressed Video Stream

针对目前大部分视频对象分割方法相当复杂而且计算量大的问题,提出了一种在压缩域粗分割,在空域精细分割的方法。该方法利用压缩域中运动向量进行聚类,得到运动对象的初始分割。将分割模板通过运动参数映射到参考帧I帧,解码初始分割区域进行Canny边缘俭测和边缘跟踪,即可得到精确的对象轮廓.该方法使得处理的数据量保持最小,节约了处理时间并得到了像素级精度的分割对象。

In the light of most current segmentation algorithms are of high complexity and huge computation, one algorithm of coarse segmentation in compressed domain and refined segmentation in spatial domain is put forward. The initial coarse segmentation masks from the motion vectors are obtained by applying Estimation Maximum(EM) algorithm. These blocks in the masks can be decompressed to obtain the origin image and the actual edges of the objects can be extracted by applying Canny edge detection and edge tracking only in the segmented regions. By using the proposed algorithm, the amount of data needed to be processed is kept in necessarily minimal, saving the computation time as well as gaining the pixel-wise edges of the segmented objects.