Final Detectron 2 Dataset

v2

Augmented

Generated on Jul 10, 2024

Dataset Split

Train Set 95%
29790Images
Valid Set 2%
718Images
Test Set 3%
830Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 5
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Crop: 0% Minimum Zoom, 25% Maximum Zoom
Rotation: Between -15° and +15°
Shear: ±10° Horizontal, ±10° Vertical
Grayscale: Apply to 21% of images
Hue: Between -22° and +22°
Saturation: Between -30% and +30%
Brightness: Between -19% and +19%
Exposure: Between -15% and +15%
Blur: Up to 4.2px
Noise: Up to 1.92% of pixels