www.pudn.com > gmm_utilities.zip > Contents.m

% Tim Bailey's Gaussian Mixture Model and Gaussian Kernel MatLab Utilities 
% Version 1.0, 2006. 
% 
% http://www.acfr.usyd.edu.au/homepages/academic/tbailey/software/software.html 
% 
% approximate_gauss_by_gmm  - Create a set of N gmms to approximate a single Gaussian 
% approximate_gauss_by_kernels- Create a set of N kernels to approximate a single Gaussian 
% 
% covariance_intersect      - Perform the covariance intersection of two Gaussians 
% gauss_divide              - Compute a/b, where a,b are Gaussians 
% gauss_multiply            - Multiply two Gaussians returning the result and the normalising weight  
% 
% gmm_addition              - Compute c = a+b, where PDFs p(a),p(b) are gmms (equivalent to gmm_convolve) 
% gmm_conditional           -  
% gmm_convolve              - Convolve two Gaussian mixtures 
% gmm_correlate             - Cross-correlate two Gaussian mixtures 
% gmm_covariance_intersect  - "Generalised" CI for gmms 
% gmm_display_1D            -  
% gmm_display_2D_contour    - 
% gmm_distance_bayes        - Bayesian distance between two gmms - normalising constant after multiplication 
% gmm_distance_bhattacharyya- Bhattacharyya distance between two gmms (Monte Carlo) 
% gmm_distance_KLD          - Kullback-Leibler divergence between two gmms (Monte Carlo) 
% gmm_divide                - Compute a/b, where a,b are gmms 
% gmm_em                    - 
% gmm_em_auto               -  
% gmm_evaluate              - Evaluate gmm at discrete points 
% gmm_marginal              - 
% gmm_multiply              - Multiply two gmms 
% gmm_normalise             - Make integral of gmm equal to one and return normalising constant 
% gmm_reduce_merge          - Reduce number of gmm components by joining  
% gmm_reduce_truncate       - Reduce number of gmm components by eliminating those with small weights 
% gmm_remove_zeros          - 
% gmm_samples               - Generate samples from gmm 
% gmm_subtract              - Compute c = a-b, where PDFs p(a),p(b) are gmms (equivalent to gmm_correlate) 
% gmm_to_gaussian           - Compute mean and variance of gmm 
% gmm_transform             - Apply a linear transform to gmm, y = Hx 
% gmm_update                - Perform a Kalman update on a gmm PDF given a gmm likelihood 
% gmm_update_linearised     - 
%  
% kernel_convolve           - 
% kernel_distance_bayes     - 
% kernel_distance_bhattacharyya-  
% kernel_distance_KLD       - 
% kernel_evaluate           - 
% kernel_multiply           - 
% kernel_normalise          - 
% kernel_reduce_merge       - 
% kernel_reduce_truncate    - 
% kernel_samples            - 
% kernel_to_gaussian        - 
% kernel_transform          - 
% kernel_update             - 
% 
% KF_update_w               - Kalman update that also returns the normalising weight of the multiplication 
% KF_update_w_simple        - Same as KF_update_w but simpler implementation