Dataset Versions
v8
median_noisy
Generated on Jul 26, 2022
382 Total Images
View All ImagesDataset Split
Train Set 97%
369Images
Valid Set 2%
6Images
Test Set 2%
7Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 416x416
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Crop: 0% Minimum Zoom, 52% Maximum Zoom
Rotation: Between -45° and +45°
Shear: ±24° Horizontal, ±24° Vertical
Grayscale: Apply to 9% of images
Hue: Between -7° and +7°
Saturation: Between -7% and +7%
Brightness: Between -7% and +7%
Blur: Up to 1.75px
Noise: Up to 4% of pixels
Cutout: 25 boxes with 8% size each