Dataset Versions

v1

Version 1 -22k Data-

Generated on May 27, 2023

Dataset Split

Train Set 90%
20778Images
Valid Set 5%
1237Images
Test Set 4%
948Images

Preprocessing

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

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Crop: 0% Minimum Zoom, 16% Maximum Zoom
Rotation: Between -45° and +45°
Shear: ±25° Horizontal, ±33° Vertical
Brightness: Between -6% and +6%