Algorithm Based on Morphological Component Analysis and Scale-Invariant Feature Transform for Image Registration

doi:10.1007/s12204-017-1807-7

Abstract

Abstract: In this paper, we proposed a registration method by combining the morphological component analysis (MCA) and scale-invariant feature transform (SIFT) algorithm. This method uses the perception dictionaries, and combines the Basis-Pursuit algorithm and the Total-Variation regularization scheme to extract the cartoon part containing basic geometrical information from the original image, and is stable and unsusceptible to noise interference. Then a smaller number of the distinctive key points will be obtained by using the SIFT algorithm based on the cartoon part of the original image. Matching the key points by the constrained Euclidean distance, we will obtain a more correct and robust matching result. The experimental results show that the geometrical transform parameters inferred by the matched key points based on MCA+SIFT registration method are more exact than the ones based on the direct SIFT algorithm.

Key words: image registration| morphological component analysis (MCA)| scale-invariant feature transform (SIFT)| key point matching

摘要： In this paper, we proposed a registration method by combining the morphological component analysis (MCA) and scale-invariant feature transform (SIFT) algorithm. This method uses the perception dictionaries, and combines the Basis-Pursuit algorithm and the Total-Variation regularization scheme to extract the cartoon part containing basic geometrical information from the original image, and is stable and unsusceptible to noise interference. Then a smaller number of the distinctive key points will be obtained by using the SIFT algorithm based on the cartoon part of the original image. Matching the key points by the constrained Euclidean distance, we will obtain a more correct and robust matching result. The experimental results show that the geometrical transform parameters inferred by the matched key points based on MCA+SIFT registration method are more exact than the ones based on the direct SIFT algorithm.

关键词: image registration| morphological component analysis (MCA)| scale-invariant feature transform (SIFT)| key point matching

CLC Number:

TN 911

WANG Gang* (王刚), LI Jingna (李京娜), SU Qingtang (苏庆堂)，ZHANG Xiaofeng (张小峰), L¨U Gaohuan (吕高焕), WANG Honggang (王洪刚). Algorithm Based on Morphological Component Analysis and Scale-Invariant Feature Transform for Image Registration[J]. Journal of shanghai Jiaotong University (Science), 2017, 22(1): 99-106.

References 16

[1]	PENG X, XU J, ZHOU Y, et al. Highly parallel linebasedimage coding for many cores [J]. IEEE Transactionson Image Processing, 2012, 21(1): 196-206.
[2]	YANG J, YU K, GONG Y, et al. Linear spatial pyramidmatching using sparse coding for image classification[C]//Computer Vision and Pattern Recognition,2009. Miami, Florida: IEEE, 2009: 1794-1801.
[3]	PAL S, MITRA M. Detection of ECG characteristicpoints using multiresolution wavelet analysis basedselective coefficient method [J]. Measurement, 2010,43(2): 255-261.
[4]	MISRA I, MOORTHI S M, DHAR D, et al. An automaticsatellite image registration technique based onHarris corner detection and Random Sample Consensus(RANSAC) outlier rejection model [C]// 2012 1stInternational Conference on Recent Advances in InformationTechnology (RAIT). India: IEEE, 2012: 68-73.
[5]	KITCHEN L, ROSENFELD A. Gray-level corner detection[R]. Maryland: Maryland Univ College ParkComputer Vision Lab, 1980.
[6]	DRESCHLER L, NAGEL H H. Volumetric model and3D trajectory of a moving car derived from monocularTV-frame sequences of a street scene [J]. ComputerGraphics and Image Processing, 1982, 20(3): 199-228.
[7]	LOWE D G. Distinctive image features from scaleinvariantkeypoints [J]. International Journal of ComputerVision, 2004, 60(2): 91-110.
[8]	ANDREU F, CASELLES V, DIAZ J I, et al. Somequalitative properties for the total variation flow [J].Journal of Functional Analysis, 2002, 188(2): 516-547.
[9]	VESE L A, OSHER S J. Modeling textures with totalvariation minimization and oscillating patterns inimage processing [J]. Journal of Scientific Computing,2003, 19(1/2/3): 553-572.
[10]	STARCK J L, ELAD M, DONOHO D. Redundantmultiscale transforms and their application for morphologicalcomponent separation [J]. Advances inImaging and Electron Physics, 2004, 132: 287-348.
[11]	STARCK J L, DONOHO D L, CAND`ES E J. Astronomicalimage representation by the curvelet transform[J]. Astronomy & Astrophysics, 2003, 398(2):785-800.
[12]	BRUCE A G, SARDY S, TSENG P. Block coordinaterelaxation methods for nonparamatric signal denoising[C]//Aerospace/Defense Sensing and Controls.USA: International Society for Optics and Photonics,1998: 75-86.
[13]	CANDES E, DEMANET L, DONOHO D, et al. Fastdiscrete curvelet transforms [J]. Multiscale Modeling &Simulation, 2006, 5(3): 861-899.
[14]	DAUGMAN J G. Uncertainty relation for resolution inspace, spatial frequency, and orientation optimized bytwo-dimensional visual cortical filters [J]. Journal ofthe Optical Society of America A: Optics Image Scienceand Vision, 1985, 2(7): 1160-1169.
[15]	MIKOLAJCZYK K, SCHMID C. A performance evaluationof local descriptors [J]. IEEE Transactionson Pattern Analysis and Machine Intelligence, 2005,27(10): 1615-1630.
[16]	SARDY S, TSENG P, BRUCE A. Robust wavelet denoising[J]. IEEE Transactions on Signal Processing,2001, 49(6): 1146-1152.