Dataset Versions

v1

2024-12-10 5:33pm

Generated on Dec 10, 2024

Dataset Split

Train Set 94%
1274Images
Valid Set 4%
51Images
Test Set 2%
25Images

Preprocessing

Auto-Orient: Applied

Augmentations

Outputs per training example: 7
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -7° and +7°
Saturation: Between -25% and +25%
Brightness: Between -3% and +3%
Blur: Up to 3.5px
Noise: Up to 0.65% of pixels