Dataset Versions

v1

2024-04-02 3:41pm

Generated on Apr 2, 2024

Dataset Split

Train Set 87%
471Images
Valid Set 9%
47Images
Test Set 4%
24Images

Preprocessing

Auto-Orient: Applied
Static Crop: 4-96% Horizontal Region, 5-95% Vertical Region
Resize: Stretch to 224x224
Auto-Adjust Contrast: Using Contrast Stretching

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Grayscale: Apply to 20% of images
Exposure: Between -15% and +15%