Dataset Versions

v6

-FINAL VERSION-

Generated on Aug 19, 2024

Dataset Split

Train Set 94%
51432Images
Valid Set 4%
2304Images
Test Set 2%
1168Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Tile: 2 rows x 2 columns
Modify Classes: 4 remapped, 0 dropped

Augmentations

Outputs per training example: 6
Saturation: Between -30% and +30%
Brightness: Between -25% and +25%
Exposure: Between -15% and +15%
Blur: Up to 4px
Noise: Up to 1.45% of pixels