IUG

dice-detector

Instance Segmentation

dice-detector Image Dataset

v4

2023-09-04 10:13am

Generated on Sep 4, 2023

Dataset Split

Train Set 95%
270Images
Valid Set 2%
5Images
Test Set 4%
10Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Grayscale: Apply to 100% of images
Hue: Between -137° and +137°
Saturation: Between -94% and +94%
Brightness: Between -70% and +70%
Exposure: Between -44% and +44%
Blur: Up to 2px
Noise: Up to 25% of pixels
Mosaic: Applied
Bounding Box: Blur: Up to 20px