How to get high accurate result with OpenCV Face Recognition Package?

I am using OpenCV Face Recognition package and Haar Cascade classifier in python. Training is done with 10 images for each individual and a total of around 60 individual’s images have been trained. Though the algorithm works, at times the results are incorrect

  1. Getting low accurate and mismatch results when face captured by camera with low light
    or spectacles.
  2. Sometimes returning unknown results for trained individual.

What can be done to improve the accuracy and prevent the mismatches ?

just curious, – those classes are building a database for identification. is this really your use-case ? what are you trying to solve here ?

We are trying build face recognition based attendance management system using opencv.
We have trained model with 10 images of each individuals(total of 60) but at times we are getting wrong results(match with wrong photo).