When you have a lot of images and want to filter them into different folders, this repo can help you.
It is an effective tool to get classification dataset. If you want to classify images using text file, check this repo.
--input
direct to the directory of your images.
--output
direct to the directory to save outcome.
--type
write down your classes with comma separation (w/o space).
i.e. --type apple,banana,melon
python3 main.py --input /media/huijo/SSD_Storage/cropInstance/ecf/raw --output /media/huijo/SSD_Storage/cropInstance/ecf/filtered --type basil,error
After applying object detection or instant segmentation, each object/instance can be cropped using bbox
.
Each bbox can be further filtered (more specified classes or false positive samples) using this repo.
This is my exaple of cropping samples.