Dataset Versions
v4
2023-04-11 5:52pm
Generated on Apr 11, 2023
2930 Total Images
View All ImagesDataset Split
Train Set 70%
2048Images
Valid Set 20%
589Images
Test Set 10%
293Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Grayscale: Applied
Augmentations
Outputs per training example: 1
Flip: Horizontal
Rotation: Between -15° and +15°
Shear: ±15° Horizontal, ±15° Vertical
Brightness: Between 0% and +20%
Blur: Up to 0.25px
Noise: Up to 3% of pixels