That would be great! Any idea how to locate the text on the image? Since face recognition can be done "online" ( http://auduno.github.io/clmtrackr/examples/facesubstitution.... ), finding text should also be possible.