This Article 
 Bibliographic References 
 Add to: 
Generalized Projection-Based M-Estimator
Dec. 2012 (vol. 34 no. 12)
pp. 2351-2364
S. Mittal, Dept. of Stat., Columbia Univ., New York, NY, USA
S. Anand, Dept. of Electr. & Comput. Eng., Rutgers Univ., Piscataway, NJ, USA
P. Meer, Dept. of Electr. & Comput. Eng., Rutgers Univ., Piscataway, NJ, USA
We propose a novel robust estimation algorithm - the generalized projection-based M-estimator (gpbM), which does not require the user to specify any scale parameters. The algorithm is general and can handle heteroscedastic data with multiple linear constraints for single and multicarrier problems. The gpbM has three distinct stages - scale estimation, robust model estimation, and inlier/outlier dichotomy. In contrast, in its predecessor pbM, each model hypotheses was associated with a different scale estimate. For data containing multiple inlier structures with generally different noise covariances, the estimator iteratively determines one structure at a time. The model estimation can be further optimized by using Grassmann manifold theory. We present several homoscedastic and heteroscedastic synthetic and real-world computer vision problems with single and multiple carriers.
Index Terms:
estimation theory,computer vision,computer vision problems,generalized projection-based M-estimator,robust model estimation algorithm,gpbM,heteroscedastic data,linear constraints,scale estimation,inlier-outlier dichotomy,noise covariances,Grassmann manifold theory,Estimation,Noise measurement,Robustness,Computational modeling,Robust estimation,Covariance matrix,RANSAC,Generalized projection-based M-estimator,robust estimation,heteroscedasticity
S. Mittal, S. Anand, P. Meer, "Generalized Projection-Based M-Estimator," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 34, no. 12, pp. 2351-2364, Dec. 2012, doi:10.1109/TPAMI.2012.52
Usage of this product signifies your acceptance of the Terms of Use.