Dataset Versions

v5

v5

Generated on Dec 2, 2024

Dataset Split

Train Set 87%
768Images
Valid Set 8%
74Images
Test Set 4%
36Images

Preprocessing

Auto-Orient: Applied
Resize: Fill (with center crop) in 150x500
Auto-Adjust Contrast: Using Histogram Equalization
Filter Null: Require at least 10% of images to contain annotations.

Augmentations

Outputs per training example: 3
Saturation: Between -25% and +25%

Similar Projects

See More