Dataset Versions
Versions
v1
2023-07-10 5:54pm
Generated on Jul 10, 2023
3462 Total Images
View All ImagesDataset Split
Train Set 88%
3030Images
Valid Set 8%
288Images
Test Set 4%
144Images
Preprocessing
Auto-Orient: Applied
Resize: Fill (with center crop) in 1920x1080
Auto-Adjust Contrast: Using Adaptive Equalization
Modify Classes: 14 remapped, 0 dropped
Augmentations
Outputs per training example: 3
Crop: 0% Minimum Zoom, 25% Maximum Zoom
Rotation: Between -40° and +40°
Shear: ±45° Horizontal, ±45° Vertical
Blur: Up to 2px