paraConfig_RaF.m

 % script: trackparam.m
%     loads data and initializes variables
%

% Copyright (C) Jongwoo Lim and David Ross.
% All rights reserved.

% DESCRIPTION OF optIONS:
%
% Following is a description of the options you can adjust for
% tracking, each proceeded by its default value.  For a new sequence
% you will certainly have to change p.  To set the other options,
% first try using the values given for one of the demonstration
% sequences, and change parameters as necessary.
%
% p = [px, py, sx, sy, theta]; The location of the target in the first
% frame.
% px and py are th coordinates of the centre of the box
% sx and sy are the size of the box in the x (width) and y (height)
%   dimensions, before rotation
% theta is the rotation angle of the box
%
% 'numsample',1000,   The number of samples used in the condensation
% algorithm/particle filter.  Increasing this will likely improve the
% results, but make the tracker slower.
%
% 'condenssig',0.01,  The standard deviation of the observation likelihood.
%
% 'affsig',[4,4,.02,.02,.005,.001]  These are the standard deviations of
% the dynamics distribution, that is how much we expect the target
% object might move from one frame to the next.  The meaning of each
% number is as follows:
%    affsig(1) = x translation (pixels, mean is 0)
%    affsig(2) = y translation (pixels, mean is 0)
%    affsig(3) = x & y scaling
%    affsig(4) = rotation angle
%    affsig(5) = aspect ratio
%    affsig(6) = skew angle
% clear all
% dataPath = 'D:\Dropbox\dropbox\Tracking\data\';
% dataPath = 'F:\dropbox\Tracking\data\';
% title = 'woman';

% switch (title)
% case 'davidin';  p = [158 106 62 78 0]; %0.9
%     opt = struct('numsample',1000, 'affsig',[4, 4,.005,.00,.001,.00], 'updateThres', 0.9);
% case 'trellis';  p = [200 100 45 49 0]; %0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.00, 0.00, 0.00, 0.0]);
% case 'car4';  p = [123 94 107 87 0]; %0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.02,.0,.001,.00]);
% case 'car11';  p = [88 139 30 25 0]; %0.8
%     opt = struct('numsample',1000,'affsig',[4,4,.005,.0,.001,.00]);
% case 'animal'; p = [350 40 100 70 0]; %0.8
%     opt = struct('numsample',1000,'affsig',[12, 12,.005, .0, .001, 0.00]);
% case 'shaking';  p = [250 170 60 70 0];% 0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.005,.00,.001,.00]);
% case 'singer1';  p = [100 200 100 300 0]; %0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.01,.00,.001,.0000]);
% case 'bolt';  p = [292 107 25 60 0]; %0.9
%     opt = struct('numsample',1000, 'affsig',[4,4,.005,.000,.001,.000], 'updateThres', 0.9);
% case 'woman';  p = [222 165 35 95 0.0]; %0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.005,.000,.001,.000]);               
% case 'bird2';  p = [116 254 68 72 0.0]; % 0.8
%     opt = struct('numsample',1000, 'affsig',[4,4,.005,.000,.001,.000]); 
% case 'surfer';  p = [286 152 32 35 0.0]; %0.8
%     opt = struct('numsample',1000,'affsig',[8,8,.01,.000,.001,.000]);     
% otherwise;  error(['unknown title ' title]);
% end

% The number of previous frames used as positive samples.
opt.maxbasis = 100;
opt.updateThres = 0.85;
opt.numsample = 400;

opt.affsig = [8, 8, 0.03, 0.005, 0.005, 0.005];
opt.normalWidth = 320;
opt.normalHeight = 240;
opt.condenssig = 0.01;
opt.tmplsize = [32, 32];
opt.retrainThres = 0.2;