HiLO

This is the official repo for "High-Level Object Fusion for Autonomous Driving using Transformers".

arXiv | DOI | BibTeX

Installation

Tested on Ubuntu 22.04 with python 3.10.18, pytorch 2.7.1 and cuda 12.8. See the exported conda environment at environment.yaml.

With conda

conda create -n hilo python==3.10
conda activate hilo

conda install cuda -c nvidia/label/cuda-12.8.1

git clone https://github.com/rst-tu-dortmund/HiLO.git
cd HiLO

pip install -r requirements.txt
pip install -e .

Other

1. Install CUDA on your machine or inside your environment Linux / Windows.
2. Install requirements and HiLO:

pip install -r requirements.txt
pip install -e .

Data

t.b.n.

Results

Source Domain	F1 @ 0.5	F1 @ 0.5	F1 @ 0.5	mAP	mAP	mAP	Checkpoint
	urban	highway	combined	urban	highway	combined
Urban	59.14	64.97	62.06	23.84	16.88	24.69	urban_hilo_max_F1
Highway	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n
Combined	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n	t.b.n

Getting Started

Define your environment as a new yaml config at src/hilo/config/environment. Make sure to set all fields as defined in base_environment.yaml.

Usage

Training

To train the model, see the following example command with experiment train/hilo_urban which trains HiLO on urban data:

python src/train.py environment=YOUR_ENVIRONMENT +experiment=train/hilo_urban

Evaluation

To evaluate the model, see the following example command with experiment eval/urban_trained/on_urban which evaluates the urban trained HiLO model on urban data:

python src/eval.py environment=YOUR_ENVIRONMENT +experiment=eval/urban_trained/on_urban

By default, this will evaluate the model on the validation split. Add this argument to compute the metrics on the test split:

+split=test

Citation

If you use this code for your research, please cite the following paper:

@INPROCEEDINGS{osterburg2025hilo,
  author={Osterburg, Timo and Albers, Franz and Diehl, Christopher and Pushparaj, Rajesh and Bertram, Torsten},
  booktitle={2025 IEEE Intelligent Vehicles Symposium (IV)}, 
  title={HiLO: High-Level Object Fusion for Autonomous Driving Using Transformers}, 
  year={2025},
  volume={},
  number={},
  pages={209-214},
  doi={10.1109/IV64158.2025.11097611}}