Dataset Versions

v2

2024-01-27 7:55am

Generated on Jan 27, 2024

Dataset Split

Train Set 94%
23709Images
Valid Set 4%
968Images
Test Set 2%
486Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 7
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Saturation: Between -25% and +25%
Brightness: Between -15% and +15%
Exposure: Between -10% and +10%
Blur: Up to 2px
Noise: Up to 0.5% of pixels