Dataset Versions

v7

2024-12-13 4:25pm

Generated on Dec 13, 2024

Dataset Split

Train Set 81%
2040Images
Valid Set 9%
232Images
Test Set 9%
232Images

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