The app keeps hitting the Rekognition API quota, and doing detection in a library would make the responses happen much faster. Suggested libraries (based on a couple hours of research):
will keep adding to this list
Using Docker might make installation a lot easier:
I haven't! Would happily accept a pull request for anything. [Edit: Previous comment was deleted...?]
Extra points for OpenCV:
maybe you can use this:
I know it not ruby but maybe reconding in nodejs would lead to a faster application
Most of the actual detection code for the libraries above is in C/C++, but I'm open to anything at this point!
Note that any new detection code would be added to https://github.com/afeld/face_detect.
@afeld I've been working on a small node.js face detection app and have a docker container built if you think it might be of use. https://github.com/neufeldtech/funnyface.
Faced looks really promising and since its using OpenCV this should be incredibly fast :D