*************************************************************************

                Department of Mathematical Sciences                       
                   The Johns Hopkins University                           
                                                                        
                             SEMINAR                                
                                                                        
*************************************************************************

Professor David Donoho                         Friday, April 27, 2000
Department of Statistics                       ***NOTE: 104 Maryland Hall
Stanford University                            Coffee: 9:30 a.m.
                                               Seminar: 10:00 a.m.

*************************************************************************

           BEYOND WAVELETS: RIDGELETS, CURVELETS, BEAMLETS

*************************************************************************

                            ABSTRACT

In my talk, I will describe some recent work with Emmanuel Candes
(CalTech) and Xiaoming Huo (Georgia Tech) motivated by the hope of building
new multiscale representations of image data.  These representations
respond to the need to deal efficiently with edges and other singularities
in images.  While these new representations have applications in
mathematical analysis and in image representation, I will stress here the
theory and applications in removing random noise from images and in
detecting very faint structures in noisy background.

*************************************************************************