Dataset Versions

v2

2024-11-09 10:00am

Generated on Nov 9, 2024

Dataset Split

Train Set 91%
618Images
Valid Set 8%
57Images
Test Set 1%
6Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Rotation: Between -7° and +7°
Shear: ±3° Horizontal, ±3° Vertical
Grayscale: Apply to 15% of images
Brightness: Between -15% and +15%
Exposure: Between -15% and +15%