Dataset Versions

v3

2023-09-21 1:20pm

Generated on Sep 21, 2023

Dataset Split

Train Set 82%
4761Images
Valid Set 9%
522Images
Test Set 9%
503Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Brightness: Between -25% and +25%
Blur: Up to 2.5px
Noise: Up to 5% of pixels