RegionInput

The names of the tracks that specify a region of the video to encode.

The track(s) you specify must contain records with UUIDData and RegionData field types.

For example, to create an event clip for each person who appears in a video you might configure a task as follows.

[EventClip]
Type=EventClip
OutputPath=...

// Track that contains video frames to encode
ImageInput=Default_Image

// Track that supplies events, so that Media Server knows
// when to start and stop the clip
EventTrack=FaceDetect.Result

This task would output a clip that contains the whole of each video frame, so there might be more than one person who appears in the output video. You can add the RegionInput parameter to the task, so that the video frames are cropped around the face of the person described by the event.

// Region of video frames to use
RegionInput=FaceDetect.Data

OpenText recommends that when you set this parameter, you also set the parameter VideoSize. All of the cropped frames are rescaled to match the size given by the VideoSize parameter. Set VideoSize=copy to rescale all of the images to the largest size of the tracked object. For example, if a person walks towards the camera they will appear larger in size. The frames of the person when they are further away are increased in size to match, so that all the video frames have a consistent size.

To add padding to the region contained in the records, so that you can see some of the surrounding area, use the Border parameter.

Type: String (comma-separated list)
Default: The video clip is not cropped to a region
Required: No
Configuration Section: TaskName
Example: RegionInput=FaceDetect.Data
See Also:

Border

EventTrack

ImageInput

VideoSize