武大&上交发布首篇「图像匹配」大领域综述!涵盖8个子领域,汇总近20年经典方法
极市导读
武汉大学和上海交通大学近日联合发布了首篇图像匹配大领域综述:《Image Matching from Handcrafted to Deep Features: A Survey》,引用文献500+,涵盖特征匹配、图匹配、点集配准等8个子领域,是一篇非常全面的大框架图像匹配综述。论文现已被IJCV2020接收。>>>极市七夕粉丝福利活动:炼丹师们,七夕这道算法题,你会解吗?
一、问题定义及分类
二、研究背景及意义
三、特征匹配研究现状
在进行特征匹配之前,我们首先需要从两幅图像中提取显著并且具有可区分性和可匹配性的点结构。常见的点结构一般为图像内容中的角点、交叉点、闭合区域中心点等具有一定物理结构的点,而提取点结构的一般思想为构建能够区分其他图像结构的响应函数 [15,32](Response Function)或者从特征线或轮廓中进行稀疏采样 [21]。为此,Morevec [19]于 1977 年首次提出了“兴趣点”的概念,并介绍了一种基于局部像素灰度差异的特征点检测方法。然而该方法存在方向、尺度、仿射和噪声上的敏感性,以及较大的时间需求。为此,大量研究者针对该问题提出了一系列的改进措施,其中著名 的 Harris [14] 角点检测器便是运用二阶矩或自相关矩阵来加速局部极值搜索并且保证方向的不变性,为了进一步减少导数的计算,一种基于局部区域像素灰度比较的快速特征提取方法被广泛应用于具有实时要求的视觉任务中,其中包括 SUSAN 算子 [32], 以及采取不同像素比较方法和比较范围的 FAST [16] 及其改进形式如:FAST-ER [33]、 AGAST[34] 等,同时还包括在实时视觉任务中应用极为广泛的 ORB 特征 [28,35]。基于像素比较的特征提取方法也称为二值特征,通常具有极高的提取效率并具有一定的方向不变性以及所提取的特征点具有较高的重复率,对后续的匹配具有重要意义,然而这类方法受尺度和仿射变换的影响较大。针对上述问题,带有尺度信息的斑点特征成为特征提取的另一种形式,其最早是由Lindeberg 等人[36] 提出的高斯拉普拉斯(Laplace of Gaussian,LoG)函数响应来实现,并从中提出了尺度空间理论,其利用高斯响应函数的圆对称性和对局部团结构的极值响应特性以及对噪声抑制能力,通过不同高斯标准差实现在尺度空间上的极值搜索,从而提取对尺度、方向和噪声鲁棒的特征点并得到相应的尺度信息。为了避免大量的计算,D.Lowe 等人 [37,38] 介绍了一种高斯差分(Difference-of-Gaussian,DoG) 法来近似 LoG 的计算,并提出了著名的 SIFT 特征描述子。基于相同的思想,Bay 等 人 [39] 在 Hessian 矩阵的基础上结合箱式滤波以及图像积分对梯度进行快速计算,提出了SURF 算子,极大程度地提升了斑点特征的检测速度。此外,许多基于 SIFT 和 SURF 的改进方法也相继被提出,其中包括减少计算量、提升仿射鲁棒性等 [40–43]。为满足精确的匹配要求,所提取的特征通常需要精确的位置信息并保证两个点集具有较高的可重复性和可匹配性。因此,大多特征提取方法中均会采用非极大值抑制(NMS)来提升局部特征点的显著性和稳定性,并且通过像素空间的插值方法估计特征点在亚像素空间的精确极值位置,具体的特征提取相关综述请参考 [5,13,44–47]。一旦两个可匹配的点集提取完成,图像匹配任务便转化为对两个特征点集进行配对。对此,目前已涌现出了许多开创性的工作及其后续的改进方案,主要从特征匹配的本质属性入手,从不同角度对特征匹配进行定义与假设,并结合相关技术手段对问题建模与求解。根据现有文献以及相关研究成果,特征匹配问题主要从直接和间接求解两个思路进行。直接匹配的思想主要是将特征匹配问题抽离为两个点集对应的问题,直接从中估计正确的点点对应关系,而间接匹配一般先通过特征点的局部描述子的相 似程度建立初步的对应关系,然后根据几何约束剔除误匹配。此外,由于深度学习 [48] (Deep Learning,DL)技术在深层特征层面强大的学习与表达能力,基于深度卷积网 络的特征匹配技术也得到了广泛关注 [4,49,50],为解决图像匹配问题提供了一个新的方向。本文将对上述解决特征匹配的技术路线中主要方法进行分析总结。3.1 直接匹配策略
3.2 间接匹配策略
3.3深度学习策略
四、特征匹配发展趋势
特征匹配问题由来已久,理论上的突破使得现有的方法具有一定的实际应用能力,然而面对诸多方面的应用需求,以及特征匹配问题本身的复杂特性,其依然是一个具有理论研究意义和实际应用价值的开放性话题,因此需要进一步地深入研究,同时深度学习技术的强大能力也使得特征匹配问题面临着进一步的突破。接下来,综合当前研究现状以及相关难题,特征匹配技术的发展趋势主要涉及以下几个方面:传统方法的进一步推进
深度学习方法的引入
协同匹配与增量匹配
[1] 马颂德. 计算机视觉: 计算理论与算法基础. 北京: 科学出版社 [M]. 1998.
[2] 宋智礼. 图像配准技术及其应用的研究 [D]. 复旦大学, 2010.
[3] 马佳义. 基于非参数模型的点集匹配算法研究 [D]. 华中科技大学, 2014.
[4] 于伟. 基于卷积神经网络特征的图像匹配研究 [D]. 哈尔滨工业大学, 2017.
[5] Zitova B, Flusser J. Image registration methods: a survey[J]. Image and vision computing, 2003, 21(11): 977–1000.
[6] Dawn S, Saxena V, Sharma B. Remote sensing image registration techniques: A survey[C]. International Conference on Image and Signal Processing. 2010: 103– 112.
[7] Pratt W K. Digital image processing john wiley & sons[J]. Inc., New York, 1991.
[8] Viola P, Wells iii W M. Alignment by maximization of mutual information[J]. International journal of computer vision, 1997, 24(2): 137–154.
[9] Barnea D I, Silverman H F. A class of algorithms for fast digital image registration[J]. IEEE transactions on Computers, 1972, 100(2): 179–186.
[10] Bracewell R N, Bracewell R N. The Fourier transform and its applications: Vol 31999[M]. [S.l.]: McGraw-Hill New York, 1986.
[11] De castro E, Morandi C. Registration of translated and rotated images using finite Fourier transforms[J]. IEEE Transactions on pattern analysis and machine intelligence, 1987(5): 700–703.
[12] Chen Q-S, Defrise M, Deconinck F. Symmetric phase-only matched filtering of Fourier-Mellin transforms for image registration and recognition[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 1994(12): 1156–1168.
[13] Tuytelaars T, Mikolajczyk K, others. Local invariant feature detectors: a survey[J]. Foundations and trends® in computer graphics and vision, 2008, 3(3): 177–280.
[14] Harris C G, Stephens M, others. A combined corner and edge detector.[C]. Alvey vision conference: Vol 15. 1988: 10–5244.
[15] BeaudetPR.Rotationallyinvariantimageoperators[C].Proc.4thInt.JointConf. Pattern Recog, Tokyo, Japan, 1978. 1978.
[16] Rosten E, Drummond T. Machine learning for high-speed corner detection[C]. European conference on computer vision. 2006: 430–443.
[17] Yi K M, Trulls E, Lepetit V, et al. Lift: Learned invariant feature transform[C]. European Conference on Computer Vision. 2016: 467–483.
[18] Matas J, Chum O, Urban M, et al. Robust wide-baseline stereo from maximally stable extremal regions[J]. Image and vision computing, 2004, 22(10): 761–767.
[19] Moravec H P. Techniques towards automatic visual obstacle avoidance[J], 1977.
[20] Brogefors G. Hierarchical chamfer matching: A parametric edge matching algorithm[J].IEEETransactionsonPatternAnalysis&MachineIntelligence,1988(6): 849–865.
[21] Belongie S, Malik J, Puzicha J. Shape matching and object recognition using shapecontexts[J].IEEETransactionsonPatternAnalysis&MachineIntelligence, 2002(4): 509–522.
[22] Mikolajczyk K, Schmid C. Indexing based on scale invariant interest points[C]. null. 2001: 525.
[23] Radke R J, Andra S, Al-kofahi O, et al. Image change detection algorithms: a systematic survey[J]. IEEE transactions on image processing, 2005, 14(3): 294– 307.
[24] Zheng L, Yang Y, Tian Q. SIFT meets CNN: A decade survey of instance retrieval[J]. IEEE transactions on pattern analysis and machine intelligence, 2018, 40(5): 1224–1244.
[25] Ma J, Ma Y, Li C. Infrared and visible image fusion methods and applications: A survey[J]. Information Fusion, 2019, 45: 153–178.
[26] Fuentes-pacheco J, Ruiz-ascencio J, Rendón-mancha J M. Visual simultaneous localization and mapping: a survey[J]. Artificial Intelligence Review, 2015, 43(1):
[27] Fan B, Kong Q, Wang X, et al. A Performance Evaluation of Local Features for Image Based 3D Reconstruction[J]. arXiv preprint arXiv:1712.05271, 2017.
[28] Mur-artal R, Montiel J M M, Tardos J D. ORB-SLAM: a versatile and accurate monocular SLAM system[J]. IEEE transactions on robotics, 2015, 31(5): 1147– 1163.
[29] Mur-artalR, TardósJD.Orb-slam2: Anopen-sourceslamsystemformonocular, stereo, andrgb-dcameras[J].IEEETransactionsonRobotics, 2017, 33(5): 1255– 1262.
[30] Hua Z, Li Y, Li J. Image stitch algorithm based on SIFT and MVSC[C]. 2010 Seventh International Conference on Fuzzy Systems and Knowledge Discovery: Vol 6. 2010: 2628–2632.
[31] Wang C, Wang L, Liu L. Progressive mode-seeking on graphs for sparse feature matching[C]. European Conference on Computer Vision. 2014: 788–802.
[32] SmithSM,BradyJM.SUSAN—anewapproachtolowlevelimageprocessing[J]. International journal of computer vision, 1997, 23(1): 45–78.
[33] Rosten E, Porter R, Drummond T. Faster and better: A machine learning approach to corner detection[J]. IEEE transactions on pattern analysis and machine intelligence, 2010, 32(1): 105–119.
[34] MairE,HagerGD,BurschkaD,etal.Adaptiveandgenericcornerdetectionbased on the accelerated segment test[C]. European conference on Computer vision. 2010: 183–196.
[35] Rublee E, Rabaud V, Konolige K, et al. ORB: An efficient alternative to SIFT or SURF.[C]. ICCV: Vol 11. 2011: 2.
[36] Lindeberg T. Feature detection with automatic scale selection[J]. International journal of computer vision, 1998, 30(2): 79–116.
[37] LoweDG, others.Objectrecognitionfromlocalscale-invariantfeatures.[C].iccv: Vol 99. 1999: 1150–1157.
[38] Lowe D G. Distinctive image features from scale-invariant keypoints[J]. International journal of computer vision, 2004, 60(2): 91–110.
[39] Bay H, Tuytelaars T, Van gool L. Surf: Speeded up robust features[C]. European conference on computer vision. 2006: 404–417.
[40] Morel J-M, Yu G. ASIFT: A new framework for fully affine invariant image comparison[J]. SIAM journal on imaging sciences, 2009, 2(2): 438–469.
[41] Abdel-hakim A E, Farag A A. CSIFT: A SIFT descriptor with color invariant characteristics[C]. 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06): Vol 2. 2006: 1978–1983.
[42] Ke Y, Sukthankar R, others. PCA-SIFT: A more distinctive representation for local image descriptors[J]. CVPR (2), 2004, 4: 506–513.
[43] Agrawal M, Konolige K, Blas M R. Censure: Center surround extremas for realtime feature detection and matching[C]. European Conference on Computer Vision. 2008: 102–115.
[44] Schmid C, Mohr R, Bauckhage C. Evaluation of interest point detectors[J]. International Journal of computer vision, 2000, 37(2): 151–172.
[45] Mukherjee D, Wu Q J, Wang G. A comparative experimental study of image feature detectors and descriptors[J]. Machine Vision and Applications, 2015, 26(4): 443–466.
[46] Uchida Y. Local feature detectors, descriptors, and image representations: A survey[J]. arXiv preprint arXiv:1607.08368, 2016.
[47] Krig S. Interest point detector and feature descriptor survey[G]. Computer vision metrics. [S.l.]: Springer, 2016: 187–246.
[48] Lecun Y, Bengio Y, Hinton G. Deep learning[J]. nature, 2015, 521(7553): 436.
[49] Choy C B, Gwak J, Savarese S, et al. Universal correspondence network[C]. Advances in Neural Information Processing Systems. 2016: 2414–2422.
[50] Rocco I, Cimpoi M, Arandjelović R, et al. Neighbourhood consensus networks[C]. Advances in Neural Information Processing Systems. 2018: 1658–1669.
[51] Bookstein F L. Principal warps: Thin-plate splines and the decomposition of deformations[J]. IEEE Transactions on pattern analysis and machine intelligence,1989, 11(6): 567–585.
[52] AradN, DynN, ReisfeldD,etal.Imagewarpingbyradialbasisfunctions: Application to facial expressions[J]. CVGIP: Graphical models and image processing, 1994, 56(2): 161–172.
[53] Chui H, Rangarajan A. A new point matching algorithm for non-rigid registration[J]. Comput. Vis. Image Understand., 2003, 89: 114–141.
[54] Myronenko A, Song X. Point Set Registration: Coherent Point Drift[J]. IEEE Trans. Pattern Anal. Mach. Intell., 2010, 32(12): 2262–2275.
[55] Cook D J, Holder L B. Mining graph data[M]. [S.l.]: John Wiley & Sons, 2006.
[56] BabaiL.Groups,Graphs,Algorithms: TheGraphIsomorphismProblem[J].Proc. Internat. Congr. of Mathematicians 2018, 2018.
[57] Yan J, Cho M, Zha H, et al. Multi-graph matching via affinity optimization with graduated consistency regularization[J]. IEEE Trans. Pattern Anal. Mach. Intell., 2016, 38(6): 1228–1242.
[58] Yan J, Wang J, Zha H, et al. Consistency-driven alternating optimization for multigraph matching: a unified approach.[J]. IEEE Trans. Image Process., 2015, 24(3): 994–1009.
[59] Umeyama S. An eigendecomposition approach to weighted graph matching problems[J]. IEEE transactions on pattern analysis and machine intelligence, 1988, 10(5): 695–703.
[60] Leordeanu M, Hebert M. A Spectral Technique for Correspondence Problems Using Pairwise Constraints[C]. Proc. IEEE Int. Conf. Comput. Vis.. 2005: 1482– 1489.
[61] Liu H, Yan S. Common Visual Pattern Discovery via Spatially Coherent Correspondence[C]. Proc. IEEE Conf. Comput. Vis. Pattern Recognit.. 2010: 1609– 1616.
[62] SuhY,ChoM,LeeKM.Graphmatchingviasequentialmontecarlo[C].European Conference on Computer Vision. 2012: 624–637.
[63] Cho M, Lee J, Lee K M. Reweighted random walks for graph matching[C]. European conference on Computer vision. 2010: 492–505.
[64] CaelliT,KosinovS.Aneigenspaceprojectionclusteringmethodforinexactgraph matching[J]. IEEE transactions on pattern analysis and machine intelligence, 2004, 26(4): 515–519.
[65] Dong J, Soatto S. Domain-size pooling in local descriptors: DSP-SIFT[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015: 5097–5106.
[66] Calonder M, Lepetit V, Strecha C, et al. Brief: Binary robust independent elementary features[C]. European conference on computer vision. 2010: 778–792.
[67] Leutenegger S, Chli M, Siegwart R Y. BRISK: Binary robust invariant scalable keypoints[M]. [S.l.]: IEEE, 2011.
[68] Alahi A, Ortiz R, Vandergheynst P. Freak: Fast retina keypoint[C]. 2012 IEEE Conference on Computer Vision and Pattern Recognition. 2012: 510–517.
[69] Johnson A E, Hebert M. Using spin images for efficient object recognition in cluttered 3D scenes[J]. IEEE Transactions on pattern analysis and machine intelligence, 1999, 21(5): 433–449.
[70] ZaharescuA,BoyerE,VaranasiK,etal.Surfacefeaturedetectionanddescription with applications to mesh matching[C]. Proc. IEEE Conf. Comput. Vis. Pattern Recognit., 2009. 2009: 373–380.
[71] FischlerMA,BollesRC.Randomsampleconsensus: aparadigmformodelfitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981, 6(6): 381–395.
[72] Torr P H, Zisserman A. MLESAC: A New Robust Estimator with Application to Estimating Image Geometry[J]. Comput. Vis. Image Understand., 2000, 78(1): 138–156.
[73] Chum O, Matas J. Matching with PROSAC - Progressive Sample Consensus[C]. CVPR. 2005: 220–226.
[74] Sattler T, Leibe B, Kobbelt L. SCRAMSAC: Improving RANSAC’s efficiency with a spatial consistency filter[C]. Proc. IEEE Int. Conf. Comput. Vis.. 2009:2090–2097.
[75] Raguram R, Chum O, Pollefeys M, et al. USAC: a universal framework for random sample consensus[J]. IEEE transactions on pattern analysis and machine intelligence, 2013, 35(8): 2022–2038.
[76] LiX,HuZ.RejectingMismatchesbyCorrespondenceFunction[J].Int.J.Comput. Vis., 2010, 89(1): 1–17.
[77] TikhonovAN, ArseninVI.Solutionsofill-posedproblems: Vol14[M].[S.l.]: Vh Winston, 1977.
[78] Boyd S, Vandenberghe L. Convex optimization[M]. [S.l.]: Cambridge university press, 2004.
[79] Ma J, Zhao J, Tian J, et al. Robust Estimation of Nonrigid Transformation for Point Set Registration[C]. CVPR. 2013: 2147–2154.
[80] Ma J, Zhao J, Tian J, et al. Regularized vector field learning with sparse approximation for mismatch removal[J]. Pattern Recognit., 2013, 46(12): 3519–3532.
[81] MaJ,ZhaoJ,TianJ,etal.RobustPointMatchingviaVectorFieldConsensus[J]. IEEE Trans. Image Process., 2014, 23(4): 1706–1721.
[82] Ma J, Zhao J, Ma Y, et al. Non-rigid visible and infrared face registration via regularized Gaussian fields criterion[J]. Pattern Recognit., 2015, 48(3): 772–784.
[83] Ma J, Qiu W, Zhao J, et al. Robust L2E Estimation of Transformation for NonRigid Registration[J]. IEEE Trans. Signal Process., 2015, 63(5): 1115–1129.
[84] WangG, WangZ, ChenY,etal.Arobustnon-rigidpointsetregistrationmethod basedonasymmetricgaussianrepresentation[J].Comput.Vis.ImageUnderstand., 2015, 141: 67–80.
[85] Wang G, Wang Z, Chen Y, et al. Context-Aware Gaussian Fields for Non-rigid Point Set Registration[C]. CVPR. 2016: 5811–5819.
[86] Wang G, Zhou Q, Chen Y. Robust Non-Rigid Point Set Registration Using Spatially Constrained Gaussian Fields[J]. IEEE Trans. Image Process., 2017, 26(4): 1759–1769.
[87] Ma J, Zhao J, Jiang J, et al. Non-Rigid Point Set Registration with Robust Transformation Estimation under Manifold Regularization[C]. Proc. AAAI Conf. Artificial Intelligence. 2017: 4218–4224.
[88] Ma J, Zhou H, Zhao J, et al. Robust Feature Matching for Remote Sensing Image Registration via Locally Linear Transforming[J]. IEEE Trans. Geosci. Remote Sens., 2015, 53(12): 6469–6481.
[89] Ma J, Zhao J, Guo H, et al. Locality preserving matching[C]. Proc. Int. Joint Conf. Artif. Intell.. 2017: 4492–4498.
[90] Ma J, Jiang J, Zhou H, et al. Guided locality preserving feature matching for remote sensing image registration[J]. IEEE Trans. Geosci. Remote Sens., 2018.
[91] BianJ, LinW-Y, MatsushitaY,etal.Gms: Grid-basedmotionstatisticsforfast, ultra-robust feature correspondence[C]. Proc. IEEE Conf. Comput. Vis. Pattern Recognit.. 2017: 2828–2837.
[92] Lin W-Y, Wang F, Cheng M-M, et al. CODE: Coherence based decision boundariesforfeaturecorrespondence[J].IEEETrans.PatternAnal.Mach.Intell.,2018, 40(1): 34–47.
[93] Lin W-Y, Cheng M-M, Lu J, et al. Bilateral functions for global motion modeling[C]. Proc. Eur. Conf. Comput. Vis.. 2014: 341–356.
[94] Simo-serra E, Trulls E, Ferraz L, et al. Discriminative learning of deep convolutionalfeaturepointdescriptors[C].ProceedingsoftheIEEEInternationalConference on Computer Vision. 2015: 118–126.
[95] Mishchuk A, Mishkin D, Radenovic F, et al. Working hard to know your neighbor’s margins: Local descriptor learning loss[C]. Advances in Neural Information Processing Systems. 2017: 4826–4837.
[96] Wei X, Zhang Y, Gong Y, et al. Kernelized subspace pooling for deep local descriptors[C].ProceedingsoftheIEEEConferenceonComputerVisionandPattern Recognition. 2018: 1867–1875.
[97] Ono Y, Trulls E, Fua P, et al. LF-Net: learning local features from images[C]. Advances in Neural Information Processing Systems. 2018: 6237–6247.
[98] DetoneD,MalisiewiczT,RabinovichA.Superpoint: Self-supervised interest point detection and description[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2018: 224–236.
[99] Han X, Leung T, Jia Y, et al. Matchnet: Unifying feature and metric learning for patch-based matching[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015: 3279–3286.
[100] WangJ,ZhouF,WenS,etal.Deep metric learning with angularl oss[C].ProceedingsoftheIEEEInternationalConferenceonComputerVision.2017: 2593–2601.
[101] Schönberger J L, Hardmeier H, Sattler T, et al. Comparative evaluation of handcrafted and learned local features[C]. Computer Vision and Pattern Recognition (CVPR), 2017 IEEE Conference on. 2017: 6959–6968.
[102] Tian Y, Fan B, Wu F, et al. L2-Net: Deep Learning of Discriminative Patch Descriptor in Euclidean Space.[C]. Cvpr: Vol 1. 2017: 6.
[103] Zbontar J, Lecun Y. Stereo matching by training a convolutional neural network to compare image patches[J]. Journal of Machine Learning Research, 2016, 17(132): 2.
[104] Revaud J, Weinzaepfel P, Harchaoui Z, et al. Deepmatching: Hierarchical deformable dense matching[J]. International Journal of Computer Vision, 2016, 120(3): 300–323.
[105] Menze M, Heipke C, Geiger A. Object Scene Flow[J]. ISPRS Journal of Photogrammetry and Remote Sensing (JPRS), 2018.
[106] ScharsteinD,SzeliskiR.ATaxonomyandEvaluationofDenseTwo-FrameStereo Correspondence Algorithms[J]. International Journal of Computer Vision, 2002, 47(1-3): 7–42.
[107] ZagoruykoS,KomodakisN.Learning to compare image patches via convolutional neural networks[C].ProceedingsoftheIEEEConferenceonComputerVisionand Pattern Recognition. 2015: 4353–4361.
[108] AltwaijryH,TrullsE,HaysJ,etal.Learning to match aerial images with deep attentive architectures[C].ProceedingsoftheIEEEConferenceonComputerVision and Pattern Recognition. 2016: 3539–3547.
[109] Balakrishnan G, Zhao A, Sabuncu M R, et al. An Unsupervised Learning Model for Deformable Medical Image Registration[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018: 9252–9260.
[110] Jiang P, Shackleford J A. CNN Driven Sparse Multi-Level B-Spline Image Registration[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018: 9281–9289.
[111] Qi C R, Su H, Mo K, et al. Pointnet: Deep learning on point sets for 3d classification and segmentation[J]. Proc. Computer Vision and Pattern Recognition (CVPR), IEEE, 2017, 1(2): 4.
[112] Deng H, Birdal T, Ilic S. Ppfnet: Global context aware local features for robust 3d point matching[J]. Computer Vision and Pattern Recognition (CVPR). IEEE, 2018, 1.
[113] YiKM,TrullsE,OnoY,etal.LearningtoFindGoodCorrespondences[C].Proc. IEEE Conf. Comput. Vis. Pattern Recognit.. 2018: 1–9.
[114] Luo Z, Shen T, Zhou L, et al. Geodesc: Learning local descriptors by integrating geometry constraints[C]. Proceedings of the European Conference on Computer Vision (ECCV). 2018: 168–183.
[115] Zhao C, Cao Z, Li C, et al. NM-Net: Mining Reliable Neighbors for Robust Feature Correspondences[J]. arXiv preprint arXiv:1904.00320, 2019.
[116] 赵键. 点模式匹配算法研究 [D]. 国防科学技术大学, 2012.
[117] 柳成荫. 基于点特征的多模与多视角图像非刚性配准算法研究 [D]. 华中科技大 学, 2018.
[118] Tron R, Zhou X, Esteves C, et al. Fast multi-image matching via density-based clustering[C]. Proceedings of the IEEE International Conference on Computer Vision. 2017: 4057–4066.
[119] Maset E, Arrigoni F, Fusiello A. Practical and efficient multi-view matching[C]. Proceedings of the IEEE International Conference on Computer Vision. 2017: 4568–4576.
[120] Hu N, Huang Q, Thibert B, et al. Distributable consistent multi-object matching[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018: 2463–2471.
推荐阅读