NBAVision

NBAVision

Object Detection

NBAVision Image Dataset

v4

2024-02-01 11:07pm

Generated on Feb 1, 2024

Popular Download Formats

Dataset Split

Train Set 87%
639Images
Valid Set 8%
61Images
Test Set 5%
37Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Flip: Horizontal
Crop: 0% Minimum Zoom, 20% Maximum Zoom
Rotation: Between -12° and +12°
Shear: ±6° Horizontal, ±6° Vertical
Grayscale: Apply to 15% of images
Hue: Between -20° and +20°
Saturation: Between -30% and +30%
Exposure: Between -15% and +15%
Cutout: 5 boxes with 2% size each
Bounding Box: Brightness: Between -15% and +15%
Bounding Box: Blur: Up to 1px