KCS5

sd_ciga

Object Detection

sd_ciga Image Dataset

v5

2023-03-18 9:16pm

Generated on Mar 18, 2023

Popular Download Formats

Dataset Split

Train Set 91%
14019Images
Valid Set 6%
877Images
Test Set 3%
444Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 500x500

Augmentations

Outputs per training example: 3
Crop: 0% Minimum Zoom, 20% Maximum Zoom
Shear: ±15° Horizontal, ±15° Vertical
Grayscale: Apply to 25% of images
Saturation: Between -25% and +25%
Cutout: 3 boxes with 10% size each
Bounding Box: 90° Rotate: Clockwise, Counter-Clockwise