Dataset Versions
v5
2024-11-14 12-19pm
Generated on Nov 14, 2024
5640 Total Images
View All Images
Dataset Split
Train Set 88%
4935Images
Valid Set 8%
470Images
Test Set 4%
235Images
Preprocessing
Auto-Orient: Applied
Resize: Fit within 256x256
Auto-Adjust Contrast: Using Adaptive Equalization
Filter Null: Require at least 80% of images to contain annotations.
Augmentations
Outputs per training example: 3
Flip: Horizontal
Crop: 0% Minimum Zoom, 20% Maximum Zoom
Shear: ±5° Horizontal, ±5° Vertical
Hue: Between -10° and +10°
Saturation: Between -10% and +10%
Brightness: Between -15% and +15%
Exposure: Between -10% and +10%
Blur: Up to 1px
Noise: Up to 0.1% of pixels