Dataset Versions
v3
2024-11-02 11:05pm
Generated on Nov 2, 2024
4531 Total Images
View All Images
Dataset Split
Train Set 88%
3966Images
Valid Set 8%
377Images
Test Set 4%
188Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 4 remapped, 1 dropped
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Grayscale: Apply to 25% of images
Hue: Between -21° and +21°
Saturation: Between -30% and +30%
Brightness: Between -24% and +24%
Exposure: Between -12% and +12%
Blur: Up to 1.5px