Abstract
Generating a more realistic 3D reconstruction point cloud is an ill-posed problem. It is a challenging task to infer 3D shape from a single image. In this paper, a two-stage training network that can reconstruct point cloud from a single image is proposed, namely, 3D-ARNet. The 3D-ARNet uses the designed image encoder with an attention mechanism to extract image features and output a simple point cloud. To improve the accuracy of point cloud reconstruction, the 3D-ARNet network contains a pre-trained point cloud auto-encoder, which a takes simple point cloud as input, and finally obtains an accurately reconstructed point cloud. The proposed approach is analyzed qualitatively and quantitatively on both synthetic and real-world datasets. Improvements are evidently demonstrated from experimental comparison results in reference to existing related state-of-the-art networks.
Similar content being viewed by others
References
Achlioptas P, Diamanti O, Mitliagkas I, Guibas LJ (2018) Learning representations and generative models for 3d point clouds. In ICML
Chang AX, Funkhouser TA, Guibas LJ, Hanrahan P, Huang QX, Li Z, Savarese S, Savva M, Song S, Su H, Xiao J, Yi L, Yu F (2015) Shapenet: An information-rich 3d model repository. ArXiv, abs/1512.03012
Choy CB, Xu D, Gwak J, Chen K, Savarese S (2016) 3d-r2n2: A unified approach for single and multi-view 3d object reconstruction. Lect Notes Comput Sci 628–644
Choi S, Nguyen AD, Kim JW, Ahn S, Lee S (2019) Point cloud deformation for single image 3d reconstruction. 2019 IEEE International Conference on Image Processing (ICIP) 2379–2383
Di X, Yu P (2017) 3d reconstruction of simple objects from a single view silhouette image 01
Fan H, Su H, Guibas L (2017) A point set generation network for 3d object reconstruction from a single image. 2017 Proc IEEE Conf Comput Vis Pattern Recognit (CVPR)
Fuentes-Pacheco J, Ruiz-Ascencio J, Rendon-Mancha JM (2015) Visual simultaneous localization and mapping: a survey. Artif Intell Rev 43(1):55–81
Gadelha M, Maji S, Wang R (2017) 3d shape induction from 2d views of multiple objects. 2017 International Conference on 3D Vision (3DV)
Girdhar R, Fouhey DF, Rodriguez M, Gupta A (2016) Learning a predictable and generative vector representation for objects. Lect Notes Comput Sci 484–499
Haming K, Peters G (2010) The structure-from-motion reconstruction pipeline - a survey with focus on short image sequences. Kybernetika 5:01
Hu J, Shen L, Sun G (2018) Squeeze-and-excitation networks
Ian J (2014) Goodfellow, Jean Pouget-Abadie, M. Aaron C. Courville, and Yoshua Bengio. Generative adversarial nets. In NIPS, Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair
Kingma DP, Ba J (2015) Adam: A method for stochastic optimization. CoRR, abs/1412.6980
Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In NIPS
Kurenkov A, Ji J, Garg A, Mehta V, Gwak J, Choy C, Savarese S (2018) Deformnet: Free-form deformation network for 3d shape reconstruction from a single image. 2018 IEEE Winter Conference on Applications of Computer Vision (WACV)
Lu Q, Xiao M, Lu Y, Yuan X, Yu Y (2019) Attention-based dense point cloud reconstruction from a single image. IEEE Access 7:137420–137431
Mandikal P, Navaneet KL, Agarwal M, Babu RV (2018) 3d-lmnet: Latent embedding matching for accurate and diverse 3d point cloud reconstruction from a single image
Navaneet KL, Mandikal P, Agarwal M, Babu RV (2019) Capnet: Continuous approximation projection for 3d point cloud reconstruction using 2d supervision. Proceedings of the AAAI Conference on Artificial Intelligence 33:8819–8826
Qi CR, Su H, Mo K, Guibas LJ (2017) Pointnet: Deep learning on point sets for 3d classification and segmentation. 2017 Proc IEEE Conf Comput Vis Pattern Recognit (CVPR)
Ramasinghe S, Khan S, Barnes N, Gould S (2019) Spectral-gans for high-resolution 3d point-cloud generation. ArXiv, abs/1912.01800
Rubner Y, Tomasi C, Guibas LJ (2000) The earth mover’s distance as a metric for image retrieval. Int J Comput Vis 40(2):99–121
Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. CoRR, abs/1409.1556
Sun R, Gao Y, Fang Z, Wang A, Zhong C (2019) Ssl-net: Point-cloud generation network with self-supervised learning. IEEE Access 7:82206–82217
Sun X, Wu J, Zhang X, Zhang Z, Zhang C, Xue T, Tenenbaum JB, Freeman WT (2018) Pix3d: Dataset and methods for single-image 3d shape modeling. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition 2974–2983
Tulsiani S, Zhou T, Efros AA, Malik J (2017) Multi-view supervision for single-view reconstruction via differentiable ray consistency. In Proc IEEE Conf Comput Vis Pattern Recognit 2626–2634
Wei Y, Liu S, Zhao W, Lu J, Zhou J (2019) Conditional single-view shape generation for multi-view stereo reconstruction. 2019 IEEE/CVF Conf Comput Vision Pattern Recognit (CVPR) 9643–9652
Wu Z, Song S, Khosla A, Yu F, Zhang L, Tang X, Xiao J (2015) 3d shapenets: A deep representation for volumetric shapes. 2015 Proc IEEE Conf Comput Vis Pattern Recognit (CVPR)
Yan X, Yang J, Yumer E, Guo Y, Lee H (2016) Perspective transformer nets: Learning single-view 3d object reconstruction without 3d supervision. ArXiv, abs/1612.00814
Zhang W, Long C, Yan Q, Xiao C (2020) Multi-stage point completion network with critical set supervision
Acknowledgements
This work was conducted during the research year of Shanghai University of Electric Power in 2020 and this work is supported by National Natural Science Foundation of China (Grant No. 51705304), Natural Science Foundation of Shanghai (Grant No. 20ZR1421300).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Chen, H., Zuo, Y. 3D-ARNet: An accurate 3D point cloud reconstruction network from a single-image. Multimed Tools Appl 81, 12127–12140 (2022). https://doi.org/10.1007/s11042-021-11433-7
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-021-11433-7