Dataset Versions

v4

2024-05-29 12:33pm

Generated on May 29, 2024

Dataset Split

Train Set 88%
8780Images
Valid Set 10%
978Images
Test Set 3%
255Images

Preprocessing

Auto-Orient: Applied

Augmentations

Outputs per training example: 2
Shear: ±30° Horizontal, ±30° Vertical
Blur: Up to 3.5px
Noise: Up to 1.99% of pixels
Cutout: 4 boxes with 15% size each