Adaptive Weighted boxes fusion

Repository based on containing Python implementation of several methods for ensembling boxes from object detection models:

• Non-maximum Suppression (NMS)
• Soft-NMS [1]
• Non-maximum weighted (NMW) [2]
• Weighted boxes fusion (WBF) [3] - new method which gives better results comparing to others

In addition to a multi-agent system (MAS) that implement WBF in a decentralized and adaptive manner.

Requirements

Python 3.*, Numpy, Numba

Examples

Pleas refer to main.py and AWBF_Examples folder (jupiter notebook for visualizing the bounding box evolution)

Description of AWBF method and citation

• https://ceur-ws.org/Vol-3813/13.pdf

If you find this code useful please cite:

@inproceedings{daoud2024introducing,
  title={Introducing Multiagent Systems to AV Visual Perception Sub-tasks: A proof-of-concept implementation for bounding-box improvement},
  author={Daoud, Alaa and Bunel, Corentin and Gu{\'e}riau, Maxime},
  booktitle={13th International Workshop on Agents in Traffic and Transportation (ATT 2024) held in conjunction with ECAI 2024},
  year={2024},
  organization={CEUR-WS}
}