Introduction
The rotate
transformation engine duplicates an existing track that contains images, and rotates the images (the input track is not modified). The engine also updates coordinates that are contained in records, so that the positions remain correct after the images have been rotated. The new track is named taskName.Output
, where taskName
is the name of the transformation task.
You might want to rotate images when:
- You want to process an image that is upside down and need to rotate the image before it is analyzed. Though you can configure some analysis engines to work regardless of orientation, you might obtain better performance by rotating the ingested image and restricting analysis to a single orientation.
- You run analysis tasks that detect faces or text in an image, and you want to make sure that copies of the image are encoded in the correct orientation.