基于时域局部线性预测的视频镜头变换检测

提出了通过时域局部线性预测，在统一框架中准确、高效地检测视频镜头的切变和叠化等镜头变换的新方法．通过理论分析和基于视频数据的线性回归，分别得到叠化预测模式和平稳镜头预测模式．通过同时比较分析两个线性预测模式的误差，在统一框架中检测切变和叠化变换．由于直接在视频流压缩域中进行分析，因此节省了视频解码的时间．采用新方法检测时仅需要对整数进行加、减、比较和移位操作，计算复杂度很低，适合在低功耗、嵌入式设备上运行．采用MPEG-7标准测试视频的实验结果显示：切变检测查全率为95．5％，查准率为93．9%；叠化变换检测查全率为82．2%，查准率为76．0％．检测一段38min的视频耗时仅3．5s．

Shot Change Detection Based on Local Linear Prediction in Time Domain

A novel approach based on local linear prediction in time domain was proposed to detect shot changes such as cuts and dissolves accurately and efficiently in a unified framework. From the theoretical analysis and linear regression of video data, two linear prediction models of dissolves and stationary scenes were developed respectively. By comparing the errors of these two models, cuts and dissolves were detected simultaneously. Since the analysis was carried out directly in the video compressed domain, lots of decode time were saved. Besides, the computational complexity was very low because only the operation of addition, subtraction, comparison, and shifting was performed on integers. Therefore, this detection is suitable for low-power devices and embedded systems. Experimental results based on the MPEG-7 standard sequences demonstrate that the recall and precision of cut detection are 95.5% and 93.9% respectively, and the recall and precision of dissolve detection are 82. 2% and 76.0% respectively. With this approach only 3.5 seconds were required for the shot change detection of a 38-minute video.