Contains code for DKM: Dense Kernelized Feature Matching for Geometry Estimation
Note The code in this repo is in active development, and the api may change without notice.
⚡ DKMv3 is out, with improved result on ScanNet1500 (+1.6 AUC@5) and MegaDepth1500 (+3.5 AUC@5)!
| @5 | @10 | @20 | |
|---|---|---|---|
| DKMv1 | 54.5 | 70.7 | 82.3 |
| DKMv2 | 56.8 | 72.3 | 83.2 |
| DKMv3 (paper) | 60.5 | 74.9 | 85.1 |
| DKMv3 (this repo) | 60.0 | 74.6 | 84.9 |
| @5 | @10 | @20 | |
|---|---|---|---|
| DKMv3 (paper) | 60.5 | 74.5 | 84.2 |
| DKMv3 (this repo) | 60.4 | 74.6 | 84.3 |
| @5 | @10 | @20 | |
|---|---|---|---|
| DKMv1 | 24.8 | 44.4 | 61.9 |
| DKMv2 | 28.2 | 49.2 | 66.6 |
| DKMv3 (paper) | 29.4 | 50.7 | 68.3 |
| DKMv3 (this repo) | 29.8 | 50.8 | 68.3 |
- Initial commit of DKMv3
- Fix compatability issues between DKM versions
- St Pauls Cathedral Benchmark
- Scannet download instructions
- Update demos for DKMv3
- Code for models can be found in dkm/models
- Code for benchmarks can be found in dkm/benchmarks
- Code for reproducing experiments from our paper can be found in experiments/
Run pip install -e .
Currently broken!
A demonstration of our method can be run by:
python demo_match.pyThis runs our model trained on mega on two images I took recently in the wild.
See Benchmarks for details.
See Training for details.
Given that the required benchmark or training dataset has been downloaded and unpacked, results can be reproduced by running the experiments in the experiments folder.
We recommend using the excellent Graph-Cut RANSAC algorithm: https://github.com/danini/graph-cut-ransac
| @5 | @10 | @20 | |
|---|---|---|---|
| DKMv3 (RANSAC) | 60.5 | 74.9 | 85.1 |
| DKMv3 (GC-RANSAC) | 65.5 | 78.0 | 86.7 |
We have used code and been inspired by https://github.com/PruneTruong/DenseMatching, https://github.com/zju3dv/LoFTR, and https://github.com/GrumpyZhou/patch2pix. We additionally thank the authors of ECO-TR for providing their benchmark.
If you find our models useful, please consider citing our paper!
@article{edstedt2022dkm,
title={DKM: Dense Kernelized Feature Matching for Geometry Estimation},
author={Edstedt, Johan and Athanasiadis, Ioannis and Wadenb{\"a}ck, M{\aa}rten and Felsberg, Michael},
journal={arXiv preprint arXiv:2202.00667},
year={2022}
}