How to evaluate on Easy/Moderate/Hard instances with Pascal VOC or Coco Evaluation metrics in Tensorflow Object Detection API?

Solution for How to evaluate on Easy/Moderate/Hard instances with Pascal VOC or Coco Evaluation metrics in Tensorflow Object Detection API?
is Given Below:

I am working with a pre-trained Faster-RCNN Inception model in the Tensorflow Object-Detection API on my own dataset. I am using Pascal-Voc evaluation metrics for evaluation, and it is working fine. However, I don’t see the option to get evaluation results with the Easy/Moderate/Hard difficulty flag.

I have difficulty flags (Easy/Moderate/Hard) available with my dataset labels. I want to integrate them with the Evaluation Metrics in Tensorflow Object Detection API.

How can I modify the evaluation script to perform metrics calculations with difficulty levels? Any suggestions would be appreciated.

Thank you so much for your help!