Pushing the Efficiency of StereoNet: Exploiting Spatial Sparsity

Georgios Zampokas, Georgios Zampokas, Christos-Savvas Bouganis, Dimitrios Tzovaras

2022

Abstract

Current CNN-based stereo matching methods have demonstrated superior performance compared to traditional stereo matching methods. However, mapping these algorithms into embedded devices, which exhibit limited compute resources, and achieving high performance is a challenging task due to the high computational complexity of the CNN-based methods. The recently proposed StereoNet network, achieves disparity estimation with reduced complexity, whereas performance does not greatly deteriorate. Towards pushing this performance to complexity trade-off further, we propose an optimization applied to StereoNet that adapts the computations to the input data, steering the computations to the regions of the input that would benefit from the application of the CNN-based stereo matching algorithm, where the rest of the input is processed by a traditional, less computationally demanding method. Key to the proposed methodology is the introduction of a lightweight CNN that predicts the importance of refining a region of the input to the quality of the final disparity map, allowing the system to trade-off computational complexity for disparity error on-demand, enabling the application of these methods to embedded systems with real-time requirements.

Paper Citation

in Harvard Style

Zampokas G., Bouganis C. and Tzovaras D. (2022). Pushing the Efficiency of StereoNet: Exploiting Spatial Sparsity. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP; ISBN 978-989-758-555-5, SciTePress, pages 757-766. DOI: 10.5220/0010919300003124

in Bibtex Style

@conference{visapp22,
author={Georgios Zampokas and Christos-Savvas Bouganis and Dimitrios Tzovaras},
title={Pushing the Efficiency of StereoNet: Exploiting Spatial Sparsity},
booktitle={Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP},
year={2022},
pages={757-766},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010919300003124},
isbn={978-989-758-555-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 5: VISAPP
TI - Pushing the Efficiency of StereoNet: Exploiting Spatial Sparsity
SN - 978-989-758-555-5
AU - Zampokas G.
AU - Bouganis C.
AU - Tzovaras D.
PY - 2022
SP - 757
EP - 766
DO - 10.5220/0010919300003124
PB - SciTePress