Up to now setting up the facer is mostly weird hacking. Most of the changes is done in the config files, but there are still situations where the code is pimped last minute. This happens especially if problems arise to produce debug output.

An integrated tool is needed that shows all the important values and parameters of face recognition and allows for setting every required config value in an easy way (adding/removing/changing person names comes to my mind). It should also allow for peeking into the program to get additional info in case something goes wrong. This starts with the ability to show the images of the raw image and the detected ROIs, possibly augmented with additional information available during image processing. This could be as simple as changing the pipeline to copy the processed image at the end of processing just before the rectangles are drawn, such that most of the time the last processed image is in the buffer.

This tool should preferably be a GUI application written using Gtkmm using the existing GUI elements and utils.