This repository holds some scripts and python notebooks that have helped me to preprocess Image data for ML training.
This script takes a video file as input and saves 1 frame per n frames. --n is a user input. Depending on the number of images required, this number can be tweaked. The default value has been set to 10.
python3 extract_frames <path-to-video> <path-to-output-folder> --n <N>
Open the preprocessing.ipynb file. This file contains code and documentation on usage.