Dataset Versions

v7

2024-01-18 1:07pm

Generated on Jan 18, 2024

Dataset Split

Train Set 86%
5445Images
Valid Set 9%
600Images
Test Set 5%
300Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -15° and +15°
Saturation: Between -34% and +34%
Noise: Up to 0.65% of pixels