Dataset Versions

v1

2024-05-29 1:28am

Generated on May 28, 2024

Dataset Split

Train Set 84%
96Images
Valid Set 14%
16Images
Test Set 2%
2Images

Preprocessing

Auto-Orient: Applied
Resize: Fit (reflect edges) in 640x480

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Grayscale: Apply to 20% of images
Hue: Between -15° and +15°
Brightness: Between -15% and +15%
Blur: Up to 1.5px