Dataset Versions
v4
2023-11-15 12:12pm
Generated on Nov 15, 2023
1696 Total Images
View All Images
Dataset Split
Train Set 86%
1452Images
Valid Set 9%
146Images
Test Set 6%
98Images
Preprocessing
Auto-Orient: Applied
Static Crop: 25-75% Horizontal Region, 0-80% Vertical Region
Resize: Stretch to 640x640
Filter Null: Require all images to contain annotations.
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise
Crop: 0% Minimum Zoom, 12% Maximum Zoom
Rotation: Between -45° and +45°
Shear: ±20° Horizontal, ±22° Vertical
Grayscale: Apply to 25% of images
Hue: Between -6° and +6°
Saturation: Between -25% and +25%
Brightness: Between -22% and +22%
Exposure: Between -10% and +10%
Blur: Up to 1.5px
Noise: Up to 2% of pixels
Cutout: 4 boxes with 2% size each