Dataset Versions

v1

2023-11-16 3:49am

Generated on Nov 15, 2023

Dataset Split

Train Set 87%
1876Images
Valid Set 9%
184Images
Test Set 4%
88Images

Preprocessing

Auto-Orient: Applied
Resize: Fill (with center crop) in 416x416
Auto-Adjust Contrast: Using Adaptive Equalization
Tile: 2 rows x 2 columns

Augmentations

Outputs per training example: 3
Bounding Box: Flip: Horizontal, Vertical
Bounding Box: 90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Bounding Box: Rotation: Between -15° and +15°
Bounding Box: Shear: ±15° Horizontal, ±15° Vertical
Bounding Box: Blur: Up to 2.5px