Dataset Versions

v1

2022-12-29 10:49pm

Generated on Dec 30, 2022

Dataset Split

Train Set 87%
276Images
Valid Set 8%
25Images
Test Set 5%
17Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Crop: 0% Minimum Zoom, 30% Maximum Zoom
Hue: Between -28° and +28°
Saturation: Between -33% and +33%
Brightness: Between -50% and +50%
Exposure: Between -40% and +40%
Noise: Up to 9% of pixels