Rating Manipulation
The emphasis is on portraits and images of models now, but I’d love to know if this is looking at content or some other data set to determine if an image has been doctored. If the latter, it would be a great tool for newsrooms to use when dealing with user generated content.