Computer Vision Dataset Processing Toolkit
This is a toolkit for processing some common computer vision datasets. This toolkit helps you, download, extract, organise, store and process the data in these datasets in a unified manner.
Note this is still a work in progress and not a completely debugged release, use with caution :)
Currently supported are:
Coming soon (not all functionality):
- OpenCV + FFMPEG
config.py contains all of the dataset paths and parameters, start by reading through that and setting up your paths
Each dataset file eg.
image/mscoco.py contains functions to download, extract, and process their corresponding data.