CLC number: TP391.41
On-line Access: 2024-08-27
Received: 2023-10-17
Revision Accepted: 2024-05-08
Crosschecked: 2020-06-04
Cited: 0
Clicked: 4993
Citations: Bibtex RefMan EndNote GB/T7714
Zhuo-hao Liu, Chang-yu Diao, Wei Xing, Dong-ming Lu. A low-overhead asynchronous consensus framework for distributed bundle adjustment[J]. Frontiers of Information Technology & Electronic Engineering, 2020, 21(10): 1442-1454.
@article{title="A low-overhead asynchronous consensus framework for distributed bundle adjustment",
author="Zhuo-hao Liu, Chang-yu Diao, Wei Xing, Dong-ming Lu",
journal="Frontiers of Information Technology & Electronic Engineering",
volume="21",
number="10",
pages="1442-1454",
year="2020",
publisher="Zhejiang University Press & Springer",
doi="10.1631/FITEE.1900451"
}
%0 Journal Article
%T A low-overhead asynchronous consensus framework for distributed bundle adjustment
%A Zhuo-hao Liu
%A Chang-yu Diao
%A Wei Xing
%A Dong-ming Lu
%J Frontiers of Information Technology & Electronic Engineering
%V 21
%N 10
%P 1442-1454
%@ 2095-9184
%D 2020
%I Zhejiang University Press & Springer
%DOI 10.1631/FITEE.1900451
TY - JOUR
T1 - A low-overhead asynchronous consensus framework for distributed bundle adjustment
A1 - Zhuo-hao Liu
A1 - Chang-yu Diao
A1 - Wei Xing
A1 - Dong-ming Lu
J0 - Frontiers of Information Technology & Electronic Engineering
VL - 21
IS - 10
SP - 1442
EP - 1454
%@ 2095-9184
Y1 - 2020
PB - Zhejiang University Press & Springer
ER -
DOI - 10.1631/FITEE.1900451
Abstract: Generally, the distributed bundle adjustment (DBA) method uses multiple worker nodes to solve the bundle adjustment problems and overcomes the computation and memory storage limitations of a single computer. However, the performance considerably degrades owing to the overhead introduced by the additional block partitioning step and synchronous waiting. Therefore, we propose a low-overhead consensus framework. A partial barrier based asynchronous method is proposed to early achieve consensus with respect to the faster worker nodes to avoid waiting for the slower ones. A scene summarization procedure is designed and integrated into the block partitioning step to ensure that clustering can be performed on the small summarized scene. Experiments conducted on public datasets show that our method can improve the worker node utilization rate and reduce the block partitioning time. Also, sample applications are demonstrated using our large-scale culture heritage datasets.
[1]Agarwal S, Mierle K, 2012. Ceres Solver. https://ceres-solver.org [Accessed on Feb. 23, 2020].
[2]Agarwal S, Snavely N, Seitz SM, et al., 2010. Bundle adjustment in the large. Proc 11th European Conf on Computer Vision, p.29-42.
[3]BocţRI, Csetnek ER, 2019. ADMM for monotone operators: convergence analysis and rates. Adv Comput Math, 45(1):327-359.
[4]Boyd S, Parikh N, Chu E, et al., 2011. Distributed optimization and statistical learning via the alternating direction method of multipliers. Found Trends Mach Learn, 3(1):1-122.
[5]Cui ZP, Tan P, 2015. Global structure-from-motion by similarity averaging. IEEE Int Conf on Computer Vision, p.864-872.
[6]Dhillon IS, 2001. Co-clustering documents and words using bipartite spectral graph partitioning. ACM Int Conf on Knowledge Discovery and Data Mining, p.269-274.
[7]Dhonju HK, Xiao W, Mills JP, et al., 2018. Share our cultural heritage (SOCH): worldwide 3D heritage reconstruction and visualization via web and mobile GIS. ISPRS Int J Geo-Inform, 7(9):360.
[8]Eriksson A, Bastian J, Chin TJ, et al., 2016. A consensus-based framework for distributed bundle adjustment. IEEE Conf on Computer Vision and Pattern Recognition, p.1754-1762.
[9]Feng J, He X, Konte B, et al., 2012. Summarization-based mining bipartite graphs. ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.1249-1257.
[10]Fuhrmann S, Langguth F, Goesele M, 2014. MVE: a multi-view reconstruction environment. EUROGRAPHICS Workshop on Graphics and Cultural Heritage, p.11-18.
[11]Furukawa Y, Curless B, Seitz SM, et al., 2010. Towards Internet-scale multi-view stereo. IEEE Computer Society Conf on Computer Vision and Pattern Recognition, p.1434-1441.
[12]Giselsson P, Boyd S, 2017. Linear convergence and metric selection for Douglas-Rachford splitting and ADMM. IEEE Trans Autom Contr, 62(2):532-544.
[13]Han JL, Shen SH, 2019. Distributed surface reconstruction from point cloud for city-scale scenes. Int Conf on 3D Vision, p.338-347.
[14]He BS, Yuan XM, 2012. On the O(1/n) convergence rate of the Douglas-Rachford alternating direction method. SIAM J Numer Anal, 50(2):700-709.
[15]Huang X, Peng YX, Yuan MK, 2020. MHTN: modal-adversarial hybrid transfer network for cross-modal retrieval. IEEE Trans Cybern, 50(3):1047-1059.
[16]Konolige K, 2010. Sparse sparse bundle adjustment. Proc British Machine Vision Conf, p.102.1-102.11.
[17]Kushal A, Agarwal S, 2012. Visibility based preconditioning for bundle adjustment. IEEE Conf on Computer Vision and Pattern Recognition, p.1442-1449.
[18]Liu ZH, Diao CY, Xing W, et al., 2019. Critical parameter consensus for efficient distributed bundle adjustment. Proc 14th Int Joint Conf on Computer Vision, Imaging and Computer Graphics Theory and Applications, p.800-807.
[19]Lourakis M, Argyros AA, 2005. Is Levenberg-Marquardt the most efficient optimization algorithm for implementing bundle adjustment? IEEE Int Conf on Computer Vision, p.1526-1531.
[20]Lourakis M, Argyros AA, 2009. SBA: a software package for generic sparse bundle adjustment. ACM Trans Math Softw, 36(1):2.
[21]Lu LP, Zhang Y, Liu K, 2019. Block partitioning and merging for processing large-scale structure from motion problems in distributed manner. IEEE Access, 7:114400-114413.
[22]Mostegel C, Rumpler M, Fraundorfer F, et al., 2016. Using self-contradiction to learn confidence measures in stereo vision. IEEE Conf on Computer Vision and Pattern Recognition, p.4067-4076.
[23]Mostegel C, Fraundorfer F, Bischof H, 2018. Prioritized multi-view stereo depth map generation using confidence prediction. ISPRS J Photogr Remote Sens, 143:167-180.
[24]Özyecşil O, Voroninski V, Basri R, et al., 2017. A survey of structure from motion. Acta Numer, 26:305-364.
[25]Peng YX, Zhai XH, Zhao YZ, et al., 2016. Semi-supervised cross-media feature learning with unified patch graph regularization. IEEE Trans Circ Syst Video Technol, 26(3):583-596.
[26]Peng YX, Huang X, Zhao YZ, 2018. An overview of cross-media retrieval: concepts, methodologies, benchmarks, and challenges. IEEE Trans Circ Syst Video Technol, 28(9):2372-2385.
[27]Rahaman H, Champion E, 2019. To 3D or not 3D: choosing a photogrammetry workflow for cultural heritage groups. Heritage, 2(3):1835-1851.
[28]Schönberger JL, Frahm JM, 2016. Structure-from-motion revisited. IEEE Conf on Computer Vision and Pattern Recognition, p.4104-4113.
[29]Schönberger JL, Zheng EL, Frahm JM, et al., 2016. Pixelwise view selection for unstructured multi-view stereo. 14th European Conf on Computer Vision, p.501-518.
[30]Shen XL, Dou Y, Mills S, et al., 2018. Distributed sparse bundle adjustment algorithm based on three-dimensional point partition and asynchronous communication. Front Inform Technol Electron Eng, 19(7):889-904.
[31]Shi JB, Malik J, 2000. Normalized cuts and image segmentation. IEEE Trans Patt Anal Mach Intell, 22(8):888-905.
[32]Simon I, Snavely N, Seitz SM, 2007. Scene summarization for online image collections. IEEE 11th Int Conf on Computer Vision, p.1-8.
[33]Snavely N, Seitz SM, Szeliski R, 2008. Skeletal graphs for efficient structure from motion. IEEE Conf on Computer Vision and Pattern Recognition, p.1-8.
[34]Triggs B, McLauchlan PF, Hartley RI, et al., 1999. Bundle adjustment—a modern synthesis. Int Workshop on Vision Algorithms, p.298-372.
[35]Wu CC, Agarwal S, Curless B, et al., 2011. Multicore bundle adjustment. IEEE Conf on Computer Vision and Pattern Recognition, p.3057-3064.
[36]Zhai XH, Peng YX, Xiao JG, 2014. Learning cross-media joint representation with sparse and semisupervised regularization. IEEE Trans Circ Syst Video Technol, 24(6):965-978.
[37]Zhang RL, Kwok JT, 2014. Asynchronous distributed ADMM for consensus optimization. Proc 31st Int Conf on Machine Learning, p.1701-1709.
[38]Zhang RZ, Li SW, Fang T, et al., 2015. Joint camera clustering and surface segmentation for large-scale multi-view stereo. Proc IEEE Int Conf on Computer Vision, p.2084-2092.
[39]Zhang RZ, Zhu SY, Shen TW, et al., 2020. Distributed very large scale bundle adjustment by global camera consensus. IEEE Trans Patt Anal Mach Intell, 42(2):291-303.
[40]Zhu SY, Fang T, Xiao JX, et al., 2014. Local readjustment for high-resolution 3D reconstruction. IEEE Conf on Computer Vision and Pattern Recognition, p.3938-3945.
[41]Zhu SY, Zhang RZ, Zhou L, et al., 2018. Very large-scale global SFM by distributed motion averaging. IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.4568-4577.
Open peer comments: Debate/Discuss/Question/Opinion
<1>