Dataset Versions

v3

Dataset without background

Generated on Sep 5, 2023

Dataset Split

Train Set 87%
450Images
Valid Set 9%
44Images
Test Set 4%
21Images

Preprocessing

Auto-Orient: Applied
Resize: Fill (with center crop) in 640x640
Modify Classes: 5 remapped, 3 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Rotation: Between -45° and +45°