Dataset Versions

v2

2024-10-20 1:10am

Generated on Oct 20, 2024

Dataset Split

Train Set 88%
303Images
Valid Set 8%
29Images
Test Set 4%
14Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Crop: 2% Minimum Zoom, 11% Maximum Zoom
Rotation: Between -3° and +3°
Shear: ±4° Horizontal, ±8° Vertical
Blur: Up to 1px