Dataset Versions

v9

2023-03-30 4:24pm

Generated on Mar 30, 2023

Dataset Split

Train Set 88%
8445Images
Valid Set 8%
764Images
Test Set 4%
421Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 3 remapped, 2 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Rotation: Between -15° and +15°
Grayscale: Apply to 50% of images
Brightness: Between -25% and +25%
Exposure: Between -25% and +25%
Noise: Up to 10% of pixels
Cutout: 5 boxes with 25% size each