Dataset Versions

v1

2024-09-02 1:32am

Generated on Sep 2, 2024

Dataset Split

Train Set 88%
5631Images
Valid Set 6%
402Images
Test Set 6%
402Images

Preprocessing

Auto-Orient: Applied
Static Crop: 0-100% Horizontal Region, 7-92% Vertical Region
Resize: Stretch to 256x256
Modify Classes: 0 remapped, 0 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Crop: 0% Minimum Zoom, 8% Maximum Zoom
Shear: ±2° Horizontal, ±2° Vertical
Noise: Up to 0.1% of pixels