Abstract
Estimating the fundamental matrix (F-matrix) is a basic problem in computer vision. The traditional algorithms are highly based on correspondences. By imprecise detecting and matching correspondences, the F-matrix is estimated incredibly. An end-to-end network (F-net) is provided in the present work without detecting and matching correspondences. To ensure estimation of an accurate F-matrix which is rank-2 with 7 degrees of freedom and scale invariance, we used the Improved convolutional block attention module (Improved-CBAM), and two self-define layers in this network. The experiments were conducted on the KITTI dataset. Two metrics, MMABS (Epipolar Constraint with Mean Absolute Value) and MMSQR (Epipolar Constraint with Mean Squared Value) were used to measure how well the epipolar constraint is satisfied by the estimated F-matrix. MMSQR and MMABS of the F-net are 0.21 and 0.11, respectively, and are 95.32 and 37.36 in the eight-point algorithm, respectively. For another end-to-end network, they are 3.48 and 2.77, respectively. F-net outperforms the other algorithms. The results demonstrated that the F-matrix can be successfully estimated by the F-net.
Similar content being viewed by others
References
Kittisak J, Parinya S, Ludmila AS, Wahidah H, Robbi R, Andino M, Abdurrahman A (2020) Pattern recognition and features selection for speech emotion recognition model using deep learning. Int J Speech Technol 23:799–806
Lei W, Juan G, Yize C, Yuanbo L, Weijie Z, Jiantao P, Hao C (2021) Automated segmentation of the optic disc from fundus images using an asymmetric deep learning network. Pattern Recogn 112(107810):1–12
Intisar RIH, Jeremiah N (2020) Deep learning approaches to biomedical image segmentation. Inform Med Unlocked 18(100297):1–12
Xulei Y, Zeng Z, Sin GT, Li W, Vijay C, Steven H (2018) Deep learning for practical image recognition:case study on Kaggle competitions. In: Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery & Data Mining ACM, pp 923–931
Xiao Y, Yuhua L, Qiang G (2021) Pipeline image diagnosis algorithm based on neural immune ensemble learning. Int J Press Vessels Pip 189(104249):1–8
Scott W, Connor G, Menghua Z, Ryan B, Nathan J (2015) Deepfocal: a method for direct focal length estimation. In: Processing of IEEE International Conference on Image Processing, pp 1369–1373
Oleksandr B, Francois R, Viktor E, Jean-Chales B (2018) Deepcalib: A deep learning Approach for Automatic Intrinsic calibration of wide field-of-view cameras. In: proceedings of European Conference on Visual Media Production, pp 1–10
Chaoning Z, Francois R, Junsik J, Dawit MA (2020) DeepPTZ: Deep self-calibration for PTZ cameras. In: Proceedings of Conference on Applications of Computer Vision, pp 1030–1038
Iaroslav M, Juha Y, Juho K, Esa R (2017) Relative camera pose estimation using convolutional neural networks. In: proceedings of Springer Conference on Advanced Concepts for Intelligent Vision Systems, pp 675–687
Zakaria L, Iaroslav M, Surya K, Juho K (2017) Camera relocalization by computing pairwise relative poses using convolutional neural network. arXiv:1707.09733
Daniel D, Tomasz M, Andrew R (2016) Deep image homography estimation. arXiv:1606.03798
Farzan EN, Robert L, Nathalie J (2017) Homography estimation from image pairs with hierarchical convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp 913–920
Daniel D, Tomasz M, Andrew R (2017) Toward geometric deep slam. arXiv:1707.07410
Ty N, Steven WC, Shreyas SS, Camillo JT, Vijay K (2018) Unsupervised deep homography: A fast and robust homography estimation model. IEEE Robot Autom Lett 3(3):2346–2353
Ignacio R, Relja A, Josef S (2017) Convolutional neural network architecture for geometric matching. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp 40–48
Xinwang L, Xinzhong Z, Miaomiao L, Lei W, En Z, Tongliang L, Marius K, Dinggang S, Jianping Y, Wen G (2020) Multiple kernel K-means with incomplete kernels. IEEE Trans Pattern Anal Mach Intell 5(42):1191–1204
Vladlen K, Rene R (2018) Deep fundamental matrix estimation. In: Proceedings of the European Conference on Computer Vision. pp1–16
Omind P, Guandao Y, Aditya P, Qiuren F, Hanqing J, Bharath H, Serge B (2018) Deep F matrix estimation without correspondences. arXiv: 1810.01575
Yesheng Z, Xu Z, Dahong Q (2020) An end to end network architecture for fundamental matrix estimation. arXiv: 2010. 15528
Tsung-Yi L, Piotr D, Ross G, Kaiming H, Bharath H, Serge B (2017) Feature pyramid networks for object detection. arXiv: 1612.03144
Sanghyun W, Jongchan P, Joon-Young L (2018) CBAM: convolutional block attention module. In: Proceedings of the European Conference on Computer Vision, pp 3–19
Andreas G, Philip L, Raquel U (2012) Are we ready for. Autonomous Driving? The KITTI vision benchmark suite. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3354–3361
Zhengyou Z (1998) Determining the epipolar geometry and its uncertainty: A review. Int J Comput Vision 27:161–195
Quan-Tuan L, Oliver DF (1996) The fundamental matrix: Theory, algorithms, and stability analysis. Int J Comput Vision 17:43–75
Lili Z, Binjie H, Jingguang Q, Manman C (2020) A deep-learning-based self-calibration time-reversal fingerprinting localization approach on Wi-Fi platform. IEEE Internet Things J 7(8):7072–7082
Zahra K, Hamid S, Ahmand RNN, Mehran M (2020) Fast, yet robust end-to-end camera pose estimation for robotic applications. Appl Intell 3:1–19
Tao Z (2021) Research on environmental landscape design based on virtual reality technology and deep learning. Microprocess Microsyst 8(103796):1–6
George F, Khalid A, Sameh Z (2021) Single-View 3D reconstruction: A Survey of deep learning methods. Computers & Graphics, p 127
Kui F, Jiansheng P, Qiwen H, Hanxiao Z (2020) Single image 3D object reconstruction based on deep learning: A review. Multimed Tools Appl 80:463–498
Mateusz M, Anders E, Engene B, Mahsa B (2020) A simple and scalable shape representation for 3D reconstruction. arXiv:2005.04623
Longuet-Higgins HC (1981) A computer algorithm for reconstructing a scene from two projections. Nature 293(5828):133–135
Ricard IR (1994) Projective reconstruction and invariants from multiple images. IEEE Trans Pattern Anal Mach Intell 16(10):1036–1040
Michel D, Marc R, Jean-Thierry L, Gerard R (1989) Determination of the attitude of 3-D objects from a single perspective view. IEEE Trans Pattern Anal Mach Intell 11(12):1265–1278
Nister D (2003) An efficient solution to the five-point relative pose problem. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 951–764
Fuchao W, Zhanyi H (2003) 5-points and 4-point algorithm to determine of the F matrix. Acta Autom Sin 29(2):175–180
Luong QT, Faugeras OD (1993) Determining the F matrix with planes: unstability and new algorithms. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 489–494
Crepey S (2003) Calibration of the local volatility in a generalized black—scholes model using Tikhonov regularization. SIAM 1 Math Anal 34(5):1183–1206
Chengyuan T, YiLeh W, Yuehhuang L (2008) F matrix estimation using evolutionary algorithms with Multi-Objective functions. J Inf Sci Eng 24(3):785–800
Wojciech C, Michel JB, Darren G, Anton VDH (2002) A new approach to constrained parameter estimation applicable to some computer vision problems. In: Proceedings of statistical methods in video processing workshop, pp 1–2
Wojciech C, Michel JB, Darren G, Anton VDH (2000) What value covariance information in estimating vision parameters? In: Proceedings of the Eighth IEEE International Conference on Computer Vision, pp 302–308
Hanzi W, David S (2004) Robust adaptive-scale parametric model estimation for computer vision. IEEE Trans Pattern Anal Mach Intell 26(11):1459–1474
Maria T, Ebroul I (2012) Improving the efficiency of a least median of squares schema for the estimation of the F matrix. Int J Pattern Recognit Artif Intell 20(05):633–648
Martin AF, Robert CB (1981) Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Commun ACM 24(6):381–385
Torr PHS, Zisserman A (2000) MLESAC: a new robust estimator with application to estimating image geometry. Comput Vis Image Underst 78(1):138–156
Torr PHS (2002) Bayesian model estimation and selection for epipolar geometry and generic manifold fitting. Int J Comput Vision 50(1):35–61
Kun Y, Rujin Z, Enhai L, Yuemao M (2019) A robust fundamental matrix estimation method based on epipolar geometric error criterion. IEEE Access 7(2019):147523-147533
Yong C, Lopez JA, Camps O, Sznaier M (2015) A convex optimization approach to robust fundamental matrix estimation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2170-2178
Fan Z, Can Z, Qi Z (2015) Method for fundamental matrix estimation combined with feature lines. Neurocomputing 160(2015): 300-307
Chunbao X, Dazheng F, Mingdong Y (2019) Soft decision optimization method for robust fundamental matrix estimation. Mach Vis Appl 30(2019):657–669
Lindeberg T (2012) Scale invariant feature transform. Scholarpedia 7(5):2012–2021
Baofeng Z, Yingkui J, Zhijun M, Yongchen L (2014) An efficient image matching method using Speed Up Robust Features. In: Proceedings of IEEE International Conference on Mechatronics & Automation, pp 553–558
Xuebing B, Jin C, Xiaokai M, Ying Z (2006) Improved feature points matching algorithm based on speed-up robust feature and oriented fast and rotated brief. J Comput Appl 36(7):1923–1926
Acknowledgements
The authors would like to express their gratitude to EditSprings (https://www.editsprings.com/) for the expert linguistic services provided.
This work is supported by the National Natural Science Foundation of China (62063034).
This work is supported by the Research and innovation project fund of Yunnan University (2020Z76).
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
Yang, R., Zhang, J. & Li, B. Estimating the fundamental matrix based on the end-to-end convolutional network. Appl Intell 52, 15517–15528 (2022). https://doi.org/10.1007/s10489-021-03103-w
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-021-03103-w