Quantum Motion Segmentation

Federica Arrigoni1   Willi Menapace2   Marcel Seelbach Benkner3   Elisa Ricci2   Vladislav Golyanik4
1Politecnico Milano   2Trent University   3University of Siegen   4Max Planck Institute for Informatics, Saarland Informatics Campus   
European Conference on Computer Vision (ECCV) 2022

Download Video: HD (MP4, 7.6 MB)

Qualitative results on sample images from the new Q-MSEG dataset (96 qubits), where each color (symbol) represents a distinct planar motion. On average, the accuracies of our QuMoSeg-v1, Mode [5], Synch [4] and Xu et al. [63] are 0.97, 0.93, 0.93 and 0.89, respectively.

Abstract

Motion segmentation is a challenging problem that seeks to identify independent motions in two or several input images. This paper introduces the first algorithm for motion segmentation that relies on adiabatic quantum optimization of the objective function. The proposed method achieves on-par performance with the state of the art on problem instances which can be mapped to modern quantum annealers.

Downloads


  • Paper, 13.6 MB

  • Source Code (github)

  • Poster
    PDF, 3.7 MB



Related Works on Quantum Computer Vision

[project page]   A Quantum Computational Approach to Correspondence Problems on Point Sets (CVPR 2020)
[project page]   Adiabatic Quantum Graph Matching with Permutation Matrix Constraints (3DV 2020)
[project page]   Quantum Permutation Synchronization (CVPR, 2021)
[project page]   Q-Match: Iterative Shape Matching via Quantum Annealing (ICCV, 2021)


Citation

BibTeX, 1 KB

 
@inproceedings{Arrigoni2022, 
authors={Federica Arrigoni and Willi Menapace and Marcel Seelbach  Benkner and Elisa Ricci and Vladislav Golyanik}, 
title = {Quantum Motion Segmentation}, 
booktitle = {European Conference on Computer Vision (ECCV)}, 
year={2022} 
}

Contact

For questions, clarifications, please get in touch with:
Federica Ariggoni federica.arrigoni@polimi.it
Vladislav Golyanik golyanik@mpi-inf.mpg.de
This page is Zotero translator friendly. Page last updated Imprint. Data Protection.