So in the same way we build pipelines to sanitise user text input (Little Bobby Tables etc) we need to treat image data in the same way - I guess a pipeline tha

So in the same way we build pipelines to sanitise user text input (Little Bobby Tables etc) we need to treat image data in the same way - I guess a pipeline that uses openCV to detect an image in full size and thumbnail - and if they are widely different flag for review ? It's still cool though