Dataset Versions
v4
2024-11-18 9:18pm
Generated on Nov 18, 2024
3710 Total Images
View All Images
Dataset Split
Train Set 88%
3255Images
Valid Set 8%
313Images
Test Set 4%
142Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Augmentations
Outputs per training example: 3
Shear: ±14° Horizontal, ±14° Vertical
Grayscale: Apply to 15% of images
Cutout: 18 boxes with 6% size each
Bounding Box: Noise: Up to 0.45% of pixels