Dataset Versions
v6
2024-11-25 7:08pm
Generated on Nov 25, 2024
25290 Total Images
View All Images
Dataset Split
Train Set 93%
23513Images
Valid Set 5%
1323Images
Test Set 2%
454Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Filter Null: Require all images to contain annotations.
Augmentations
Outputs per training example: 7
Crop: 0% Minimum Zoom, 20% Maximum Zoom
Rotation: Between -30° and +30°
Shear: ±25° Horizontal, ±25° Vertical
Grayscale: Apply to 40% of images
Hue: Between -81° and +81°
Saturation: Between -30% and +30%
Brightness: Between -49% and +49%
Exposure: Between -30% and +30%
Blur: Up to 4.5px
Noise: Up to 4.97% of pixels
Cutout: 6 boxes with 12% size each