Dataset Versions
Versions
v11
2023-05-05 11:17am
Generated on May 5, 2023
26822 Total Images
View All Images
Dataset Split
Train Set 87%
23433Images
Valid Set 8%
2223Images
Test Set 4%
1166Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 416x416
Modify Classes: 19 remapped, 4 dropped
Filter Null: Require all images to contain annotations.
Augmentations
Outputs per training example: 3
Shear: ±9° Horizontal, ±10° Vertical
Grayscale: Apply to 29% of images
Hue: Between -14° and +14°
Saturation: Between -40% and +40%
Brightness: Between -33% and +33%
Exposure: Between -30% and +30%
Noise: Up to 5% of pixels