Dataset Versions

v1

2023-11-17 12:16am

Generated on Nov 17, 2023

Dataset Split

Train Set 87%
228Images
Valid Set 8%
22Images
Test Set 4%
11Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Rotation: Between -14° and +14°
Blur: Up to 2.5px
Cutout: 13 boxes with 7% size each
Bounding Box: Rotation: Between -10° and +10°
Bounding Box: Shear: ±15° Horizontal, ±15° Vertical
Bounding Box: Noise: Up to 5% of pixels