Dataset Versions

v4

2023-02-20 9:23pm

Generated on Feb 21, 2023

Dataset Split

Train Set 82%
243Images
Valid Set 11%
33Images
Test Set 6%
19Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Filter Null: Require all images to contain annotations.

Augmentations

Outputs per training example: 3
Shear: ±15° Horizontal, ±15° Vertical
Blur: Up to 5.25px