Dataset Versions
v7
2023-11-25 3:14pm
Generated on Nov 25, 2023
5533 Total Images
View All Images
Dataset Split
Train Set 92%
5091Images
Valid Set 5%
253Images
Test Set 3%
189Images
Preprocessing
Resize: Fit (reflect edges) in 800x800
Modify Classes: 6 remapped, 8 dropped
Filter Null: Require all images to contain annotations.
Augmentations
Outputs per training example: 3
Flip: Vertical
Shear: ±10° Horizontal, ±10° Vertical
Saturation: Between -25% and +25%
Brightness: Between -25% and +25%
Exposure: Between -25% and +25%
Noise: Up to 5% of pixels
Cutout: 16 boxes with 4% size each