Dataset Versions

v3

2024-04-08 1:19pm

Generated on Apr 8, 2024

Dataset Split

Train Set 87%
5778Images
Valid Set 9%
567Images
Test Set 4%
261Images

Preprocessing

Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Contrast Stretching
Tile: 3 rows x 3 columns

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -25° and +25°
Shear: ±20° Horizontal, ±20° Vertical
Saturation: Between -30% and +30%
Brightness: Between -31% and +31%
Blur: Up to 1.5px
Noise: Up to 3.02% of pixels
Mosaic: Applied
Bounding Box: Flip: Horizontal, Vertical
Bounding Box: Rotation: Between -15° and +15°
Bounding Box: Shear: ±13° Horizontal, ±13° Vertical
Bounding Box: Brightness: Between -15% and +15%
Bounding Box: Blur: Up to 1.6px
Bounding Box: Noise: Up to 0.1% of pixels