Dataset Versions

v14

2024-08-29 3:40pm

Generated on Aug 29, 2024

Dataset Split

Train Set 88%
2043Images
Valid Set 8%
195Images
Test Set 4%
96Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Filter Null: Require at least 75% of images to contain annotations.

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Rotation: Between -15° and +15°
Brightness: Between -25% and +25%
Exposure: Between -10% and +10%
Blur: Up to 2.5px
Noise: Up to 0.5% of pixels