Weighted boxes fusion: Ensembling boxes from different object detection models

TL;DR

This paper introduces weighted boxes fusion, a new ensemble method for object detection that combines bounding box predictions from multiple models using confidence scores, leading to top challenge results.

Contribution

The paper proposes a novel weighted boxes fusion algorithm that improves object detection ensemble performance by leveraging confidence scores for bounding box averaging.

Findings

Achieved top results on Open Images and COCO datasets.

Demonstrated improved detection accuracy over individual models.

Provided publicly available source code for the method.

Abstract

In this work, we present a novel method for combining predictions of object detection models: weighted boxes fusion. Our algorithm utilizes confidence scores of all proposed bounding boxes to constructs the averaged boxes. We tested method on several datasets and evaluated it in the context of the Open Images and COCO Object Detection tracks, achieving top results in these challenges. The source code is publicly available at https://github.com/ZFTurbo/Weighted-Boxes-Fusion