Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speed

Project Page | Paper

Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like Speed
Yifan Wang^*, Xingyi He^*, Sida Peng, Dongli Tan, Xiaowei Zhou
CVPR 2024

realtime_demo.mp4

TODO List

• Inference code and pretrained models
• Code for reproducing the test-set results
• Add options of flash-attention and torch.compiler for better performance
• jupyter notebook demo for matching a pair of images
• Training code

Installation

conda env create -f environment.yaml
conda activate eloftr
pip install torch==2.0.0+cu118 --index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt 

The test and training can be downloaded by download link provided by LoFTR

We provide the our pretrained model in download link

Reproduce the testing results with pytorch-lightning

You need to setup the testing subsets of ScanNet and MegaDepth first. We create symlinks from the previously downloaded datasets to data/{{dataset}}/test.

# set up symlinks
ln -s /path/to/scannet-1500-testset/* /path/to/EfficientLoFTR/data/scannet/test
ln -s /path/to/megadepth-1500-testset/* /path/to/EfficientLoFTR/data/megadepth/test

Inference time

conda activate eloftr
bash scripts/reproduce_test/indoor_full_time.sh
bash scripts/reproduce_test/indoor_opt_time.sh

Accuracy

conda activate eloftr
bash scripts/reproduce_test/outdoor_full_auc.sh
bash scripts/reproduce_test/outdoor_opt_auc.sh
bash scripts/reproduce_test/indoor_full_auc.sh
bash scripts/reproduce_test/indoor_opt_auc.sh

Training

conda env create -f environment_training.yaml  # used a different version of pytorch, maybe slightly different from the inference environment
pip install -r requirements.txt
conda activate eloftr_training
bash scripts/reproduce_train/eloftr_outdoor.sh eloftr_outdoor

Citation

If you find this code useful for your research, please use the following BibTeX entry.

@inproceedings{wang2024eloftr,
  title={{Efficient LoFTR}: Semi-Dense Local Feature Matching with Sparse-Like Speed},
  author={Wang, Yifan and He, Xingyi and Peng, Sida and Tan, Dongli and Zhou, Xiaowei},
  booktitle={CVPR},
  year={2024}
}