Mismatch Removal Based on Gaussian Mixture Model for Aircraft Surface Texture Mapping

Aiming at the fact of lower efficiency and higher time cost for feature matching in aircraft surface texture mapping process, a novel mismatch removal method based on Gaussian mixture model is proposed to increase correct corresponding feature matching point pairs. The detection and initial point sets for corresponding pairs are carried out, and a vector field is interpolated between the two matching of ORB feature points. The Gaussian mixture model(GMM) is introduced and a prior is taken to force the smoothness of the field, which is based on the Tikhonov regularization in vector-valued reproducing kernel Hilbert space(RKHS). In order to obtain the optimal estimation, the MAP solution of a Bayesian model with latent variables, which could be performed by Expectation Maximization (EM) algorithm, is utilized to determine the correct correspondence. The experimental results show that the algorithm could remove mismatches effectively and the classification for feature points is excellent. The calculation time is greatly reduced, which enhanced real-time performance of aircraft surface texture mapping process.

[1]  Gary R. Bradski,et al.  ORB: An efficient alternative to SIFT or SURF , 2011, 2011 International Conference on Computer Vision.

[2]  Tao Song,et al.  Research on image matching based on improved RANSAC-SIFT algorithm , 2017, 2017 16th International Conference on Optical Communications and Networks (ICOCN).

[3]  Junkai Yi,et al.  Comparison and Study of Classic Feature Point Detection Algorithm , 2012, 2012 International Conference on Computer Science and Service System.

[4]  Erin P. J. Pearse,et al.  Continuum versus discrete networks, graph Laplacians, and reproducing kernel Hilbert spaces , 2019, Journal of Mathematical Analysis and Applications.

[5]  Miao Ma,et al.  Hybrid GA Variational Bayes Inference of Finite Mixture Models for Voxel Classification in Brain Images , 2010 .

[6]  Soummya Kar,et al.  Convergence Analysis of Distributed Inference with Vector-Valued Gaussian Belief Propagation , 2016, J. Mach. Learn. Res..

[7]  Petros Daras,et al.  Robust SIFT-based feature matching using Kendall's rank correlation measure , 2009, 2009 16th IEEE International Conference on Image Processing (ICIP).

[8]  Lingzhou Xue,et al.  Nonparametric Finite Mixture of Gaussian Graphical Models , 2018, Technometrics.

[9]  Marina L. Gavrilova,et al.  Improving RANSAC Feature Matching with Local Topological Information , 2012, 2012 Ninth International Symposium on Voronoi Diagrams in Science and Engineering.

[10]  Václav Skala,et al.  A New Approach to Vector Field Interpolation, Classification and Robust Critical Points Detection Using Radial Basis Functions , 2018, CSOC.

[11]  Zhuowen Tu,et al.  Robust Point Matching via Vector Field Consensus , 2014, IEEE Transactions on Image Processing.

[12]  Florence d'Alché-Buc,et al.  Input Output Kernel Regression: Supervised and Semi-Supervised Structured Output Prediction with Operator-Valued Kernels , 2016, J. Mach. Learn. Res..

[13]  Vittorio Murino,et al.  A Unifying Framework in Vector-valued Reproducing Kernel Hilbert Spaces for Manifold Regularization and Co-Regularized Multi-view Learning , 2014, J. Mach. Learn. Res..

[14]  Abhishake Rastogi,et al.  Multi-penalty regularization in learning theory , 2016, J. Complex..

[15]  Lorenzo Rosasco,et al.  Multi-output learning via spectral filtering , 2012, Machine Learning.

[16]  Zahra Hossein-Nejad,et al.  Image registration based on SIFT features and adaptive RANSAC transform , 2016, 2016 International Conference on Communication and Signal Processing (ICCSP).

[17]  Paolo Novati,et al.  Multi-parameter Arnoldi-Tikhonov methods , 2013 .

[18]  Shiliang Sun,et al.  Infinite mixtures of multivariate Gaussian processes , 2013, 2013 International Conference on Machine Learning and Cybernetics.

[19]  Shuqiang Yang,et al.  Outliers Elimination Based Ransac for Fundamental Matrix Estimation , 2013, 2013 International Conference on Virtual Reality and Visualization.

[20]  Julianne Chung,et al.  Learning regularization parameters for general-form Tikhonov , 2017 .

[21]  Bo Li,et al.  Local Feature Descriptor for Image Matching: A Survey , 2019, IEEE Access.

[22]  Thomas Sikora,et al.  Regularized Gradient Descent Training of Steered Mixture of Experts for Sparse Image Representation , 2018, 2018 25th IEEE International Conference on Image Processing (ICIP).

[23]  Lothar Reichel,et al.  Arnoldi-Tikhonov regularization methods , 2009 .

[24]  Hongke Xu,et al.  Image feature points matching via improved ORB , 2014, 2014 IEEE International Conference on Progress in Informatics and Computing.

[25]  Yue Wang,et al.  An improved RANSAC based on the scale variation homogeneity , 2016, J. Vis. Commun. Image Represent..

[26]  Zhenhao Wang,et al.  Mixed Gaussian Models for Modeling Fluctuation Process Characteristics of Photovoltaic Outputs , 2019, Front. Energy Res..

[27]  Jieping Ye,et al.  Multi-task Vector Field Learning , 2012, NIPS.

[28]  Peng Li,et al.  Variational inference with Gaussian mixture model and householder flow , 2019, Neural Networks.

[29]  Wen Gao,et al.  A new frame interpolation method with pixel-level motion vector field , 2014, 2014 IEEE Visual Communications and Image Processing Conference.