Dataset Versions

v3

2024-04-07 1:49pm

Generated on Apr 7, 2024

Dataset Split

Train Set 89%
3543Images
Valid Set 7%
279Images
Test Set 4%
144Images

Preprocessing

Auto-Orient: Applied
Resize: Fit within 640x640
Auto-Adjust Contrast: Using Adaptive Equalization

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise
Brightness: Between -18% and +18%
Blur: Up to 2px
Noise: Up to 0.1% of pixels