Dataset Versions

v5

2024-11-20 5:17pm

Generated on Nov 20, 2024

Dataset Split

Train Set 82%
99Images
Valid Set 14%
17Images
Test Set 4%
5Images

Preprocessing

Auto-Orient: Applied
Resize: Fit within 640x640

Augmentations

Outputs per training example: 3
Rotation: Between -45° and +45°
Shear: ±10° Horizontal, ±10° Vertical
Noise: Up to 0.1% of pixels