Dataset Versions

v1

2024-10-07 12:20am

Generated on Oct 7, 2024

Dataset Split

Train Set 88%
2412Images
Valid Set 8%
232Images
Test Set 4%
108Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Tile: 2 rows x 2 columns

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Noise: Up to 0.05% of pixels