Abstract
Fundamental matrix estimation for wide baseline images is significantly difficult due to the fact that the proportion of inliers in putative correspondences is generally very low. Traditional robust fundamental matrix estimation methods, such as RANSAC, will encounter the problems of computational inefficiency and low accuracy when outlier ratio is high. In this paper, a novel robust estimation method called inlier set sample optimization is proposed to solve these problems. First, a one-class support vector machine-based preselection algorithm is performed to efficiently select a candidate inlier set from putative SIFT correspondences according to distribution consistency of features in location, scale and orientation. Then, the quasi-optimal inlier set is refined iteratively by maximizing a soft decision objective function. Finally, fundamental matrix is estimated with the optimal inlier set. Experimental results show that the proposed method is superior to several state-of-the-art robust methods in speed, accuracy and stability and is applicable to wide baseline images with large differences.





Similar content being viewed by others
References
Hartley R, Zisserman A (2003) Multiple view geometry in computer vision, 2nd edn. Cambridge University Press, Cambridge
Armangué X, Salvi J (2003) Overall view regarding fundamental matrix estimation. Image Vis Comput 21:205–220
Zia MZ, Stark M, Schiele B, Schindler K (2013) Detailed 3d representations for object recognition and modelling. IEEE Trans Pattern Anal Mach Intell 35:2608–2623
Naikal N, Yang AY, Sastry SS (2011) Informative feature selection for object recognition via sparse PCA. In: International conference on computer vision, pp 818–825
Wu B, Zhang Y, Zhu Q (2011) A triangulation-based hierarchical image matching method for wide-baseline images. Photogramm Eng Remote Sens 77:695–708
Marcon M, Frigerio E, Sarti A, Tubaro S (2012) 3D wide baseline correspondences using depth-maps. Signal Process Image Commun 27:849–855
Pizarro D, Bartoli A (2012) Feature-based deformable surface detection with self-occlusion reasoning. Int J Comput Vis 97:54–70
Tola E, Lepetit V, Fua P (2010) Daisy: an efficient dense descriptor applied to wide-baseline stereo. IEEE Trans Pattern Anal Mach Intell 32:815–830
Kim D, Paik J (2010) Gait recognition using active shape model and motion prediction. IET Comput Vis 4:25–36
Lowe DG (2004) Distinctive image features from scale-invariant keypoints. Int J Comput Vis 60:91–110
Bay H, Ess A, Tuytelaars T, Van Gool L (2008) Speeded-up robust features (SURF). Comput Vis Image Underst 110:346–359
Wang K, Xiao P, Feng X, Wu G (2011) Image feature detection from phase congruency based on two-dimensional Hilbert transform. Pattern Recognit Lett 32:2015–2024
Yang W, Sun C, Zhang L (2011) A multi-manifold discriminant analysis method for image feature extraction. Pattern Recognit 44:1649–1657
Muja M, Lowe DG (2009) Fast approximate nearest neighbors with automatic algorithm configuration. VISAPP 1:331–340
Chen HY, Lin YY, Chen BY (2013) Robust feature matching with alternate hough and inverted hough transforms. In: IEEE conference on computer vision and pattern recognition, pp 2762–2769
Miksik O, Mikolajczyk K (2012) Evaluation of local detectors and descriptors for fast feature matching. In: International conference on pattern recognition, pp 2681–2684
Sharma K, Kim SG, Singh MP (2012) An improved feature matching technique for stereo vision applications with the use of self-organizing map. Int J Precis Eng Manuf 13:1359–1368
Fischler M, Bolles R (1981) Random sample consensus: a paradigm for model fitting with application to image analysis and automated cartography. Commun ACM 24:381–395
Torr PH, Murray DW (1997) The development and comparison of robust methods for estimating the fundamental matrix. Int J Comput Vis 24:271–300
Torr PH, Zisserman A (2000) MLESAC: a new robust estimator with application to estimating image geometry. Comput Vis Image Underst 78:138–156
Torr PHS (2002) Bayesian model estimation and selection for epipolar geometry and generic manifold fitting. Int J Comput Vis 50:35–61
Chum O, Matas J, Kittler J (2003) Locally optimized RANSAC. In: Michaelis B, Krell G (eds) Pattern recognition. Springer, Berlin, pp 236–243
Shi XB, Liu F, Wang Y et al (2011) A Fundamental matrix estimation algorithm based on point weighting strategy. In: Proceedings of international conference on virtual reality and visualization. IEEE computer society press, Washington DC, pp 24–29
Tordoff BJ, Murray DW (2005) Guided-MLESAC: faster image transform estimation by using matching priors. IEEE Trans Pattern Anal Mach Intell 27:1523–1535
Chum O, Matas J (2005) Matching with PROSAC-progressive sample consensus. In: IEEE conference on CVPR, vol 1, pp. 220–226
Xu M, Lu J (2012) Distributed RANSAC for the robust estimation of three-dimensional reconstruction. IET Comput Vis 6:324–333
Adam A, Rivlin E, Shimshoni I (2001) ROR: rejection of outliers by rotations. IEEE Trans Pattern Anal Mach Intell 23:78–84
Zhang D, Wang Y, Tao W (2012) Epipolar geometry estimation for wide baseline stereo. Int J Precis Eng Manuf 2:38–45
Zhang K, Li X, Zhang J (2014) A robust point-matching algorithm for remote sensing image registration. Geosci Remote Sens Lett 11:469–473
Zhou HB, Zhang D, Chen C et al (2011) Discarding wide baseline mismatches with global and local transformation consistency. Electron Lett 47:25–26
Choi S, Kim T, Yu W (2009) Performance evaluation of RANSAC family. In: Proceedings of the British machine vision conference (BMVC), pp 1–12
Moisan L, Stival B (2004) A probabilistic criterion to detect rigid point matches between two images and estimate the fundamental matrix. Int J Comput Vis 57:201–218
Schölkopf B, Platt JC, Shawe-Taylor J et al (2001) Estimating the support of a high-dimensional distribution. Neural Comput 13:1443–1471
Zhang Z (1998) Determining the epipolar geometry and its uncertainty: a review. Int J Comput Vis 27:161–195
Zhao F, Wang H, Chai X, Ge S (2009) A fast and effective outlier detection method for matching uncalibrated images. In: International conference on image processing, pp. 2097–2100
Mills S (2013) Relative orientation and scale for improved feature matching. In: International conference on image processing, pp 3484–3488
Ni K, Jin H, Dellaert F (2009) GroupSAC: efficient consensus in the presence of groupings. In: International conference on computer vision, pp 2193–2200
Hempstalk K, Frank E, Witten IH (2008) One-class classification by combining density and class probability estimation. In: Machine learning and knowledge discovery in databases. Springer, Berlin, pp. 505–519
Bartkowiak AM (2011) Anomaly, novelty, one-class classification: a comprehensive introduction. Int J Comput Syst Ind Manag Appl 3:061–071
Fathy ME, Hussein AS, Tolba MF (2011) Fundamental matrix estimation: a study of error criteria. Pattern Recognit Lett 32:383–391
Chang CC, Lin CJ (2011) LIBSVM: a library for support vector machines. ACM Trans Intell Syst Technol TIST 2:1–27
Fawcett T (2006) An introduction to ROC analysis. Pattern Recognit Lett 27:861–874
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Xiao, CB., Feng, DZ. & Yuan, MD. An efficient fundamental matrix estimation method for wide baseline images. Pattern Anal Applic 21, 35–44 (2018). https://doi.org/10.1007/s10044-016-0561-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10044-016-0561-z