Reproducing VOT2014 TPAMI paper results

You can use the script below to generate the results used in arXiv:1503.01313. Assuming that you have correctly initialized VOT2014 workspace, the script generates three separate reports for different ranking approaches that were included in the paper. We have used VOT toolkit tagged vot2014-final to generate the results. If you will use a more recent version of the toolkit be aware that some function names have changed so the script will have to be adapted.

% Initialize the workspace
[sequences, experiments] = vot_environment();

% Use all VOT2014 trackers (the file should contain identifiers of all trackers)
trackers = tracker_load('vot2014.txt');

% Set raw robustness sensitivity parameter to 100
set_global_variable('report_ranking_sensitivity', 100);

% Run the per-attribute ranking analysis
set_global_variable('report_ranking_average', 'weighted_mean');
context = create_report_context('report_vot2014_challenge', 'latex', true, 'raw', true);
report_challenge(context, experiments, trackers, sequences);

% Run the per-sequence pooled analysis
set_global_variable('report_ranking_average', 'pool');
context = create_report_context('report_vot2014_sequence_pool', 'latex', true, 'raw', true);
report_article(context, experiments, trackers, sequences, 'arplot', true); 

% Run the per-sequence weighted analysis
set_global_variable('report_ranking_average', 'weighted_mean');
context = create_report_context('report_vot2014_sequence_weighted', 'latex', true, 'raw', true);
report_article(context, experiments, trackers, sequences, 'arplot', true);