BABULAL AGRAWAL

Wood defect detection

Instance Segmentation

Wood defect detection Image Dataset

v1

2023-11-18 2:32pm

Generated on Nov 18, 2023

Dataset Split

Train Set 98%
28779Images
Valid Set 1%
306Images
Test Set 1%
203Images

Preprocessing

Auto-Orient: Applied
Resize: Fit (reflect edges) in 1024x1024

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Crop: 0% Minimum Zoom, 50% Maximum Zoom
Rotation: Between -45° and +45°
Shear: ±45° Horizontal, ±45° Vertical
Hue: Between -180° and +180°
Saturation: Between -25% and +25%
Exposure: Between -20% and +20%
Noise: Up to 10% of pixels