Dataset Versions

v1

2024-09-10 4:49pm

Generated on Sep 10, 2024

Dataset Split

Train Set 88%
7773Images
Valid Set 8%
723Images
Test Set 4%
364Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Crop: 0% Minimum Zoom, 26% Maximum Zoom
Shear: ±14° Horizontal, ±14° Vertical
Blur: Up to 4.5px
Noise: Up to 1.05% of pixels