Dataset Versions

v3

2023-07-14 1:36pm

Generated on Jul 14, 2023

Dataset Split

Train Set 88%
4416Images
Valid Set 8%
418Images
Test Set 4%
212Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Modify Classes: 1 remapped, 1 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Saturation: Between -25% and +25%
Brightness: Between -50% and +50%
Blur: Up to 2px
Noise: Up to 10% of pixels