Dataset Versions

v2

2024-05-12 12:52pm

Generated on May 12, 2024

Dataset Split

Train Set 87%
2322Images
Valid Set 8%
218Images
Test Set 4%
114Images

Preprocessing

Auto-Orient: Applied
Resize: Fit (black edges) in 256x256
Auto-Adjust Contrast: Using Adaptive Equalization
Grayscale: Applied
Filter Null: Require at least 50% of images to contain annotations.

Augmentations

Outputs per training example: 3
Flip: Horizontal
Bounding Box: Rotation: Between -7° and +7°
Bounding Box: Blur: Up to 2px