Data Augmentation on YOLO

Data augmentation techniques:

• Translation *
• Cropping
• Noise *
• Brightness *
• Contrast *
• Saturation *
• Gaussian blur *

Build virtual environment:

python -m venv ./venv
source ./venv/bin/activate
pip install -r requirements.txt

Run

python3 main.py --images <IMAGES_FOLDER> --labels <LABELS_FOLDER> 
--output <OUTPUT_FOLDER> --nprocess <NUMBER_OF_AUGMENTED_IMAGES>

python3 main.py --images ../Videos/20231006_144024_tp00003/output/images/ --labels ../Videos/20231006_144024_tp00003/output/labels/ --output input --nprocess 2

splitfolders --ratio .8 .1 .1 -- input/