I'm testing IBM's Watson Visual Recognition using Node-RED, I've trained it to identify some elements in the image, but I wonder if it's possible to get the exact position of these elements.
At this time it is not possible to get a location for anything other than a face which you do using the detect_faces end point. However we recognize this is a valuable feature so please “stay tuned”
Update: This is now possible with the v4 "Custom Object Detection" API: https://cloud.ibm.com/docs/visual-recognition?topic=visual-recognition-getting-started-tutorial