Dataset Versions

v2

2023-05-31 11:37pm

Generated on May 31, 2023

Dataset Split

Train Set 94%
2115Images
Valid Set 5%
114Images
Test Set 1%
14Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 8 remapped, 0 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -11° and +11°
Shear: ±15° Horizontal, ±15° Vertical
Blur: Up to 3px
Noise: Up to 5% of pixels
Cutout: 3 boxes with 16% size each