Multi-class Segment Error Table (MSET) performance evaluation for continuous recognition. Basic Matlab scripts.

October 2006, by Jamie A. Ward, ETH Zurich
=========================

Scripts
--------
test_mset_plotting.m 
This gives an example of how to use the scripts. ( In the absence of any decent tutorial, probably this is the best place to start. )

mset.m:   [t, s, e] = mset(pred, ground, varargin)

mset_segments.m:  [r, lab] = mset_segments( pred, ground, varargin )

mset_print.m: texout = mset_print( nClasses, varargin )

read_seg_info.m: out = read_seg_info( segment_info, sample_rate, include_classes, seq_len, format )

write_seg_info.m: segment_info = write_seg_info( sample_rate, input, format )

plot_mset_errors.m (incomplete): M=plot_mset_errors(numArgs, chartType, varargin)