Dataset Versions
v2
2022-11-20 7:37pm
Generated on Nov 21, 2022
21679 Total Images
View All Images
Dataset Split
Train Set 91%
19809Images
Valid Set 4%
970Images
Test Set 4%
900Images
Preprocessing
Auto-Orient: Applied
Resize: Stretch to 640x640
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise
Crop: 0% Minimum Zoom, 21% Maximum Zoom
Rotation: Between -25° and +25°
Saturation: Between -50% and +50%
Brightness: Between -50% and +50%
Blur: Up to 5.25px
Noise: Up to 14% of pixels
Cutout: 2 boxes with 32% size each