I am planning to build a facial recognition system using a Raspberry Pi 3A+ and the official Raspberry Pi No-IR camera module. I've chosen the No-IR model instead of the standard model because the system I plan to build will have to work under low light conditions.
I reckon the system will have difficulties to recognize trained faces, and therefore I was thinking of applying an effect to the training images before using them to train the system, in order to simulate the same effect the No-IR camera produces. I think this will avoid issues in the facial recognition system, but I am not completely sure.
Have you ever developed such a recognition system using a No-IR camera? What considerations did you take?
Thanks in advance.