| 摘要 |
Various embodiments for performing motion detection for video processing are described. In one or more embodiments, motion detection may be performed by computing a temporal difference measurement, a local spatial complexity measurement, and a content dependent motion detection value based on the temporal difference measurement and the local spatial complexity measurement. In some implementations, the content dependent motion detection value may comprise a division of the temporal difference measurement by the local spatial complexity measurement. In other implementations, an approximated content dependent motion detection value may be computed without division to reduce computational complexity. The content dependent motion detection value may be used to perform various adaptive post-processing operations such as for de-interlacing, noise reduction, frame rate conversion, and so forth. Other embodiments are described and claimed. |
