Dataset Versions
v3
2023-08-18 10:28am
Generated on Aug 18, 2023
831 Total Images
View All ImagesDataset Split
Train Set 87%
726Images
Valid Set 6%
54Images
Test Set 6%
51Images
Preprocessing
Auto-Orient: Applied
Resize: Fill (with center crop) in 224x224
Modify Classes: 4 remapped, 0 dropped
Filter Null: Require all images to contain annotations.
Augmentations
Outputs per training example: 3
Flip: Horizontal
Rotation: Between -45° and +45°
Hue: Between -25° and +25°
Saturation: Between -40% and +40%
Brightness: Between -25% and +25%
Exposure: Between -15% and +15%
Noise: Up to 3% of pixels