Dataset Versions

v7

2024-09-03 1:06am

Generated on Sep 3, 2024

Dataset Split

Train Set 93%
3816Images
Valid Set 3%
134Images
Test Set 4%
155Images

Preprocessing

Resize: Stretch to 1024x1024
Auto-Adjust Contrast: Using Contrast Stretching
Modify Classes: 2 remapped, 0 dropped

Augmentations

Outputs per training example: 3
Rotation: Between -15° and +15°
Hue: Between -19° and +19°
Saturation: Between -25% and +25%
Brightness: Between -19% and +19%
Exposure: Between -12% and +12%
Blur: Up to 2.7px
Noise: Up to 1.65% of pixels