Tec

DripBot2.0

Object Detection

DripBot2.0 Image Dataset

v9

2023-06-14 12:40am

Generated on Jun 14, 2023

Popular Download Formats

Dataset Split

Train Set 82%
3757Images
Valid Set 13%
582Images
Test Set 6%
262Images

Preprocessing

Auto-Orient: Applied
Resize: Fit within 400x400
Auto-Adjust Contrast: Using Adaptive Equalization
Modify Classes: 23 remapped, 27 dropped
Filter Null: Require all images to contain annotations.

Augmentations

Outputs per training example: 2
Hue: Between -32° and +32°
Brightness: Between -33% and +33%
Exposure: Between -37% and +37%
Mosaic: Applied
Bounding Box: Shear: ±15° Horizontal, ±15° Vertical
Bounding Box: Noise: Up to 5% of pixels