Dataset Versions
v2
2024-08-31 6:36pm
Generated on Aug 31, 2024
3775 Total Images
View All Images
Dataset Split
Train Set 91%
3441Images
Valid Set 5%
187Images
Test Set 4%
147Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -24° and +24°
Shear: ±10° Horizontal, ±10° Vertical
Grayscale: Apply to 17% of images
Hue: Between -29° and +29°
Exposure: Between -18% and +18%
Blur: Up to 1.5px
Noise: Up to 1.44% of pixels
Cutout: 1 boxes with 10% size each