stephenrparsons
commented
9 years ago In the event that doing this would require backend processing of images, the image coordinates will likely be necessary. To get from canvas clicks to image pixel coordinates: http://stackoverflow.com/questions/18417287/openseadragon-image-cordinates
With this the user would be able to "paint" the current clipping region onto the canvas. It would stay there when they then move the mouse elsewhere or change layers in the viewer. This could ultimately be used to create a composite, "canonical" image which could have pieces from each layer.