Robust regression methods for computer vision: A review

Regression analysis (fitting a model to noisy data) is a basic technique in computer vision, Robust regression methods that remain reliable in the presence of various types of noise are therefore of considerable importance. We review several robust estimation techniques and describe in detail the least-median-of-squares (LMedS) method. The method yields the correct result even when half of the data is severely corrupted. Its efficiency in the presence of Gaussian noise can be improved by complementing it with a weighted least-squares-based procedure. The high time-complexity of the LMedS algorithm can be reduced by a Monte Carlo type speed-up technique. We discuss the relationship of LMedS with the RANSAC paradigm and its limitations in the presence of noise corrupting all the data, and we compare its performance with the class of robust M-estimators. References to published applications of robust techniques in computer vision are also given.

[1]  P. Rousseeuw.  Least Median of Squares Regression , 1984 .

[2]  Bradley Efron,et al.  Computer-Intensive Methods in Statistical Regression , 1988 .

[3]  William H. Press,et al.  Numerical recipes , 1990 .

[4]  Svetha Venkatesh,et al.  Edge detection is a projection , 1989, Pattern Recognit. Lett..

[5]  X. Zhuang,et al.  Recovering 3-D motion parameters from image sequences with gross errors , 1989, [1989] Proceedings. Workshop on Visual Motion.

[6]  Howard Wainer,et al.  Robust Regression & Outlier Detection , 1988 .

[7]  Iain M. Johnstone,et al.  The Resistant Line and Related Regression Methods , 1985 .

[8]  P. Holland,et al.  Robust regression using iteratively reweighted least-squares , 1977 .

[9]  Thomas P. Hettmansperger,et al.  Weighted least-squares rank estimates , 1983 .

[10]  Frederick Mosteller,et al.  Exploring Data Tables, Trends and Shapes. , 1986 .

[11]  G. W. Brown,et al.  On Median Tests for Linear Hypotheses , 1951 .

[12]  B. Ripley,et al.  Robust Statistics , 2018, Encyclopedia of Mathematical Geosciences.

[13]  Louis A. Jaeckel.  Estimating Regression Coefficients by Minimizing the Dispersion of the Residuals , 1972 .

[14]  Xinhua Zhuang,et al.  Pose estimation from corresponding point data , 1989, IEEE Trans. Syst. Man Cybern..

[15]  Frederick Mosteller,et al.  Data Analysis and Regression , 1978 .

[16]  R. Haralick,et al.  A facet model for image data , 1981 .

[17]  Robert M. Haralick,et al.  2D-3D pose estimation , 1988, [1988 Proceedings] 9th International Conference on Pattern Recognition.

[18]  Behrooz Kamgar-Parsi,et al.  A Nonparametric Method for Fitting a Straight Line to a Noisy Image , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[19]  Siegfried Heiler,et al.  Robust Estimates in Linear Regression - A Simulation Approach , 1981 .

[20]  H. Theil.  A Rank-Invariant Method of Linear and Polynomial Regression Analysis , 1992 .

[21]  H. Edelsbrunner,et al.  Computing Least Median of Squares Regression Lines and Guided Topological Sweep , 1990 .

[22]  Layne T. Watson,et al.  Robust window operators , 2005, Machine Vision and Applications.

[23]  L. Marton,et al.  Advances in Electronics and Electron Physics , 1958 .

[24]  J.M. Steele,et al.  Algorithms and complexity for least median of squares regression , 1986, Discret. Appl. Math..

[25]  Thomas S. Huang,et al.  The Effect of Median Filtering on Edge Estimation and Detection , 1987, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  A. Siegel.  Robust regression using repeated medians , 1982 .

[27]  Azriel Rosenfeld,et al.  Robust computer vision: a least median of squares based approach , 1989 .

[28]  Antti J. Koivo,et al.  Robust image modeling for classification of surface defects on wood boards , 1989, IEEE Trans. Syst. Man Cybern..

[29]  R. L. Kashyap,et al.  Robust Image Models and Their Applications , 1988 .

[30]  D. Ruppert.  Robust Statistics: The Approach Based on Influence Functions , 1987 .

[31]  A. R. Hanson,et al.  Robust estimation of camera location and orientation from noisy data having outliers , 1989, [1989] Proceedings. Workshop on Interpretation of 3D Scenes.

[32]  D. F. Andrews,et al.  A Robust Method for Multiple Linear Regression , 1974 .

[33]  J. Tukey,et al.  The Fitting of Power Series, Meaning Polynomials, Illustrated on Band-Spectroscopic Data , 1974 .

[34]  Edward J. Coyle,et al.  Root properties and convergence rates of median filters , 1985, IEEE Trans. Acoust. Speech Signal Process..

[35]  Robert C. Bolles,et al.  A RANSAC-Based Approach to Model Fitting and Its Application to Finding Cylinders in Range Data , 1981, IJCAI.

[36]  Robert C. Bolles,et al.  Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography , 1981, CACM.