Dataset Versions
v1
feather classify less data and use adjusted equalidation
Generated on May 25, 2023
2400 Total Images
View All Images
Dataset Split
Train Set 87%
2097Images
Valid Set 8%
197Images
Test Set 4%
106Images
Preprocessing
Auto-Orient: Applied
Static Crop: 6-91% Horizontal Region, 19-62% Vertical Region
Resize: Stretch to 640x640
Auto-Adjust Contrast: Using Adaptive Equalization
Filter Null: Require at least 15% of images to contain annotations.
Augmentations
Outputs per training example: 3
Flip: Horizontal, Vertical
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Rotation: Between -33° and +33°
Hue: Between -25° and +25°
Brightness: Between -3% and +3%
Exposure: Between -5% and +5%
Blur: Up to 1.25px
Noise: Up to 5% of pixels