ZFTurbo/Weighted-Boxes-Fusion
Library implementing Weighted Boxes Fusion and related ensembling methods to combine predictions from multiple object detection models.
Not currently ranked — collecting fresh signals.
star history
Provides Python implementations of several box ensembling techniques for object detection: Non-maximum Suppression, Soft-NMS, Non-maximum Weighted, and Weighted Boxes Fusion (WBF). These methods combine and optimize bounding box predictions from multiple models by matching boxes across models using IoU thresholds and confidence-weighted averaging. Designed to improve final detection accuracy by leveraging ensemble predictions.
Frequently asked
- What is ZFTurbo/Weighted-Boxes-Fusion?
- Library implementing Weighted Boxes Fusion and related ensembling methods to combine predictions from multiple object detection models.
- Is Weighted-Boxes-Fusion open source?
- Yes — ZFTurbo/Weighted-Boxes-Fusion is open source, released under the MIT license.
- What language is Weighted-Boxes-Fusion written in?
- ZFTurbo/Weighted-Boxes-Fusion is primarily written in Python.
- How popular is Weighted-Boxes-Fusion?
- ZFTurbo/Weighted-Boxes-Fusion has 1.8k stars on GitHub.
- Where can I find Weighted-Boxes-Fusion?
- ZFTurbo/Weighted-Boxes-Fusion is on GitHub at https://github.com/ZFTurbo/Weighted-Boxes-Fusion.