This R library provides R functions for smoothing noisy images so that edges are preserved. Some of them preserve also corners, some of them eliminate outliers, and some do both. These functions were developed in the DFG project Mu 1031/4-1/2 and are described in more detail in:

• Chu, C. K., Glad, I. K., Godtliebsen, F., Marron, J. S. (1998). Edge-preserving smoothers for image processing. J. Amer. Statist. Assoc. 93, 526-541.
• Müller, Ch.H. (2002). Robust estimators for estimating discontinuous functions.  Metrika 55, 99-109.
• Müller, Ch.H. (2002). Comparison of high-breakdown-point estimators for image denoising.  Allg. Stat. Archiv 86, 307-321.
• Hillebrand, M. and Ch.H. Müller (2002). On consistency of redescending M-kernel smoothers. Submitted to Metrika.
• Hillebrand, M. (2003). On Robust Corner-Preserving Smoothing in Image Processing. Ph.D. thesis, University of Oldenburg, Germany.

1. epsi.zip is a zip file of the window version which can be called by "packages".

2. epsi_1.0-1.tar.gz contains the original files and the library is installed under UNIX by the command "R CMD INSTALL epsi".
 

This R library provides R functions for detection of pixel positions close to edges and for finding regression clusters within these pixel positions which provides the true edges. The library provides not only functions for detection of linear edges but also for detection of circles. These functions were developed in the DFG project Mu 1031/4-1/2 and are described in more detail in:

• Qiu, P. (1997).  Nonparametric estimation of jump surface.  The Indian Journal of Statistics 59, Series A, 268-294.
• Garlipp, T. and Müller, Ch. H. (2004). Regression Clustering with Redescending M-Estimators. In: Daniel Baier and Klaus-Dieter Wernecke (eds.): Innovations in Classification, Data Science, and Information Systems.  Springer-Verlag, Heidelberg,  38-45.
• Müller, Ch.H. and T. Garlipp (2002). Simple consistent cluster methods based on redescending M-estimators with an application to edge identification in images.  To appear in: Journal of Multivariate Analysis.
• T. Garlipp (2004). On robust jump detection in regression surfaces with applications to image analysis. Ph.D. thesis, University of Oldenburg, Germany.

1. edci.zip is a zip file of the window version which can be called by "packages".

2. edci_1.0-1.tar.gz contains the original files and the library is installed under UNIX by the command "R CMD INSTALL edci".
 

This R library provides R functions for calculating the Student location-scale depth and the Student median described in the paper "Mizera, I. and Ch.H. Müller (2003). Location-scale depth. To appear in the Journal of American Statistical Association". These functions are based on Theorems 8 and 9 of this paper and do not use the transformation to the Klein disk. In particular, by using Theorem 9, an exact method for calculating the depth contours for continuous distributions is provided. Additionally, functions are included which provide the scale parameter with maximum depth for given location parameters.

1. lsdepth_R_win.zip is a zip file of the window version which can be called by "packages".

2. lsdepth_R.zip contains the original files and the library is installed under UNIX by the command "R CMD INSTALL lsdepth".