Dataset Versions
v4
2024-09-19 8:44am
Generated on Sep 19, 2024
642 Total Images
View All Images
Dataset Split
Train Set 96%
615Images
Valid Set 2%
14Images
Test Set 2%
13Images
Preprocessing
Auto-Orient: Applied
Resize: Fit (white edges) in 162x360
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -15° and +15°
Shear: ±13° Horizontal, ±13° Vertical
Brightness: Between -25% and +25%
Blur: Up to 2.5px
Noise: Up to 1.52% of pixels
Bounding Box: Brightness: Between -10% and +10%