| 摘要 |
In general, in one aspect, the frames of a received video are analyzed using traditional film mode detection (continually capturing data and comparing the data to known parameters for various film modes) until a film mode is detected. Once the film mode is detected, the frames of the received video are analyzed using mode based tracking where measurements are continually predicted based on the mode detected and averaged with actual measurements. The use of the averaged measurements limit the effect of noise captured in the actual measurements. The averaged measurements are continually compared to known parameters for the detected mode to determine if the detected mode is still valid. If the detected mode is valid, the mode based tacking is continued. If the detected mode is no longer valid (e.g., averaged measurements not correlated with known parameters), the process returns to the traditional film mode detection.
|
