Crop Transformation Engine
The crop transformation engine produces cropped images. The region data is usually supplied by an analysis task. For example, after running face detection you can use this engine to produce cropped images that show detected faces. Alternatively, you can crop images to a fixed region by setting the Region
parameter, in which case region data in the input records are ignored.
The engine also translates any co-ordinates contained in records, so that their position is correct for the cropped image. For example, face detection outputs the locations of a person's eyes. The co-ordinates are modified so that the positions are correct for the cropped image.
Configuration Parameter | Description |
---|---|
Border | The amount of padding to add to the supplied region, before cropping the image. |
Input | The name of the track that contains the images to crop, with region data. |
Region | Specifies the region to use when cropping the image. Set this parameter only when you want to specify a fixed region. |
Type | The type of transformation engine to use. Set this parameter to crop . |
This transformation engine produces a single output track, named taskName.Output
, where taskName
is the name of the transformation task.