Dataset Versions

v2

2024-04-25 10:03am

Generated on Apr 25, 2024

Dataset Split

Train Set 96%
4791Images
Valid Set 4%
176Images
Test Set 0%
20Images

Preprocessing

Auto-Orient: Applied
Static Crop: 3-95% Horizontal Region, 10-90% Vertical Region
Resize: Stretch to 1080x640

Augmentations

Outputs per training example: 3
Grayscale: Apply to 15% of images
Exposure: Between -17% and +17%
Blur: Up to 2.5px
Noise: Up to 0.12% of pixels
Bounding Box: Shear: ±10° Horizontal, ±10° Vertical