Dataset Versions

v2

2024-06-09 7:57am

Generated on Jun 8, 2024

Dataset Split

Train Set 88%
867Images
Valid Set 8%
83Images
Test Set 4%
40Images

Preprocessing

Auto-Orient: Applied
Resize: Fill (with center crop) in 640x640

Augmentations

Outputs per training example: 3
Grayscale: Apply to 21% of images
Hue: Between -56° and +56°
Noise: Up to 2.04% of pixels
Cutout: 9 boxes with 8% size each