Xiaoming Huo's Software Web Page

This page contains software that can reproduce most of results in the papers (figures, tables, etc) by my collaborators and I. Most of the papers mentioned below can be downloaded at my publication website.


·         HRM contains Matlab codes to implement the algorithm in “A Hessian regularized nonlinear time series model” by Jie Chen and X.  Huo, which is to appear in the Journal of Computational and Graphical Statistics.

·         AddiTest Package realizing branch-and-bound subset search, additional optimality test, etc. It was used to produce the three examples in the paper: "New Optimality Test for Branch-and-Bound Based Comprehensive Subset Search," by Xuelei (Sherry) Ni and Xiaoming Huo. The paper was first submitted in August 2005.

·         MDFS Multiscale Detection Of Filamentary Structures software. This package realize the algorithm proposed in the paper: "Adaptive Multiscale Detection Of Filamentary Structures In A Background Of Uniform Random Points," by Ery Arias-Castro, David L. Donoho and Xiaoming Huo, Annals of Statistics, 2006.

·         JBEAM software. This includes a beamlet coder for curves, which is the first faithfully progressive coder (that we know of) for curves. It can be used to regenerate results in "JBEAM: multiscale curve coding via beamlets" by Huo and Chen, which will appear in IEEE Trans. Image Processing. (added in Jan. 2005)

·         FBP software. This includes a novel tree pruning method that was introduced in "A frontier-based tree pruning algorithm and automatic identification of inadmissibility" by Huo, Tsui, Kim, and Wang, which will appear in INFORMS Journal on Computing. The advantages of this method include a more faithful realization of Cross-Validation principle in building a tree model. (added in Jan. 2005)

·         Dynamic Programming for High Cencentration Regions. This software has been used in paper "A network flow approach in finding maximum likelihood estimate of high concentration regions" by Huo and Lu, Computational Statistics and Data Analysis, 46 (1): 33-56, 28 May. (added in Jan. 2005)

·         CTD Dynamic programming. This program utilizes dynamic programming approaches to solve a wide spectrum of Connect-The-Dots (CTD) problems, which are fully described and discussed in "Dynamic programming methods for `connecting the dots' in scattered point sets" by Huo, Donoho, Tovey, and Arias-Castro. (added in Jan. 2005)

·         Cascade experiments. It reproduces the results in "Building a cascade detector and applications in automatic target recognition" by Huo and Chen. Appeared in Applied Optics: Information Processing, 43 (2): 293-303, 10 January, 2004. We studied a new way to build a tree classifier: cascade. (added in Jan. 2005)

·         Tree based subsequence modelling experiments. We explored the idea of using tree models to efficiently find sparse events in a sequence. (added in Jan. 2005)


Before 2003

·         WaveLab. When I was a graduate student, I had been actively involved in this project. It contains MATLAB based implementation of wavelet transforms, and many other related algorithms, with a lot of demos. Users can read the MATLAB source codes to learn how to reproduce computational results that have been reported in many papers. See the WaveLab website for more information. (added on 01/20/02)

·         A beamlet and wedgelet toolbox. BMIA Browser is a software package that allows users to regenerate figures and computational results that have been used in the paper, "Beamlets and Multiscale Image Analysis" by Donoho and Huo. Now it has been integrated into BeamLab. Go to my publication page to download the paper in PDF. (added on 01/22/02)