run_RPT.m

%
%   Reliable Patch Trackers: Robust Visual Tracking by Exploiting Reliable Patches
%
%   Yang Li, 2015
%   http://ihpdep.github.io
%
%   This is the research code of our RPT tracker. You can check the
%   details in our CVPR paper.
%

function results = run_rpt(seq, res_path, bSaveImage)


	%default settings
	 kernel_type = 'linear'; 
	 feature_type = 'hog'; 



	%parameters according to the paper. at this point we can override
	%parameters based on the chosen kernel or feature type
	kernel.type = kernel_type;
	
	features.gray = false;
	features.hog = false;
	    features.hogcolor = false;
        
	padding = 3;  %extra area surrounding the target
	lambda = 1e-4;  %regularization
	output_sigma_factor = 0.1;  %spatial bandwidth (proportional to target)
	
	switch feature_type
	case 'gray',
		interp_factor = 0.075;  %linear interpolation factor for adaptation

		kernel.sigma = 0.2;  %gaussian kernel bandwidth
		
		kernel.poly_a = 1;  %polynomial kernel additive term
		kernel.poly_b = 7;  %polynomial kernel exponent
	
		features.gray = true;
		cell_size = 1;
		
	case 'hog',
		interp_factor = 0.02;
		
		kernel.sigma = 0.5;
		
		kernel.poly_a = 1;
		kernel.poly_b = 9;
		
		features.hog = true;
		features.hog_orientations = 9;
		cell_size = 4;
		     case 'hogcolor',
		interp_factor = 0.02;
		
		kernel.sigma = 0.5;
		
		kernel.poly_a = 1;
		kernel.poly_b = 9;
		
		features.hogcolor = true;
		features.hog_orientations = 9;
		cell_size = 4;
	otherwise
		error('Unknown feature.')
	end
	%running in benchmark mode - this is meant to interface easily
		%with the benchmark's code.
		
		%get information (image file names, initial position, etc) from
		%the benchmark's workspace variables
		seq = evalin('base', 'subS');
		target_sz = seq.init_rect(1,[4,3]);
		pos = seq.init_rect(1,[2,1]) + floor(target_sz/2);
		img_files = seq.s_frames;
		video_path = [];
		
		%call tracker function with all the relevant parameters
		[positions,targetSize, time] = tracker(video_path, img_files, pos, target_sz, ...
			padding, kernel, lambda, output_sigma_factor, interp_factor, ...
			cell_size, features, 0);
		
		%return results to benchmark, in a workspace variable
		rects = [positions(:,2) - targetSize(:,2)/2, positions(:,1) - targetSize(:,1)/2];
		rects(:,3) = targetSize(:,2);
		rects(:,4) = targetSize(:,1);
		results.type = 'rect';
		results.res = rects;
        
        fps = numel(img_files) / time;

		fprintf(' FPS:% 4.2f\n', fps);

	results.fps = fps ;
	results.type='rect';
end