Quantum Robust Fitting

Paper

Tat-Jun Chin ¹, David Suter ², Shin-Fang Chng ¹, James Quach ³.

¹ Australian Institute for Machine Learning (AIML), University of Adelaide, ² School of Computing and Security, Edith Cowan University ³ School of Physical Sciences, The University of Adelaide

About

This is the official repository for Quantum Robust Fitting.

Getting Started

This demo runs in MATLAB, and has been tested on macOS Catalina and Ventura.

Dependencies

• gurobi (https://www.gurobi.com) (optional)

Demo 1: 2D Line Fitting

This demo demonstrates an example of computing the exact influence on the 2D line fitting problem.

1. Run main.m in demo_influence folder.

Demo 2: Homography Estimation

This demo demonstrates an example of computing the exact influence on solving a homography instance for 20 feature correspondences.

1. Run main.m in demo_homography_small folder.

Demo 3: Large-Scale Homography Estimation

This demo demonstrates an example of computing the approximate influence on solving a large-scale homography instance for more than 200 feature correspondences.

1. Run main.m in demo_homography_large folder.

Reproducibility

We provide the results of large-scale homography estimation for a church instance. If you wish to plot the results, please follow the steps below:

1. Download the results from the following link: https://drive.google.com/drive/folders/1_Z5Be2T78u2PQfWPL0_bl1uvUZlCfDMY?usp=sharing

   Please note that you will find two results at the provided link: accv_official.zip contains the influence corresponding to the results in the paper, and run2.zip contains the influence for a recent run.

2. Place the data in demo_homography_large/output directory.

3. Run evaluate_approx_influence.m. You will then obtain the following plots,

   

👩‍💻 Citation

This code is for non-commercial use. If you find our work useful in your research, please consider citing our paper:

 @inproceedings{chin2020quantum,
 title={Quantum robust fitting},
 author={Chin, Tat-Jun and Suter, David and Ch'ng, Shin-Fang and Quach, James},
 booktitle={Proceedings of the Asian Conference on Computer Vision},
 year={2020}
 }