Course information:
    Intelligent Multimedia Systems
    Practial assignments

Collaborators:
    F. Huizinga [0418862] - folkerthuizinga@gmail.com
    B. Stoeller [0426857] - astone.git@gmail.com

--------------------------------------------------------------------------------

TRACKER

NAME
    tracker - Mean shift tracker implementation

SYNOPSIS
    tracker VIDEO_FILE
    tracker VIDEO_FOLDER

DESCRIPTION
	The mean shift tracker tracks an object in a sequence of video frames. You
	specify a path to a video file or a path to a folder containg separate
	video frames as images.

OPTIONS
    In the settings file globals.m several options can be set. Be aware of the
    fact that the file globals.mat can override the setting in global.m. This
    file is generated by testall.m. It should automatically be removed by
    testall.m but it might still exist if testall.m is preemptively terminated.

    DEBUG - Debug mode
        Specifies the debug mode. The debug mode can be a boolean, a string or
        a set of strings.    
            false- display only the most important debug information
            true - display any debug information
            IMAQ - Image aquisition
            TGSL - Target selection
            KRNL - Kernel calculations
            HIST - Histogram calculations
            TEST - Test cycles

        e.g. {'IMAQ', 'TGSL'} displays all debug messages from both image
        aquisition and target selection. Note that some important debug messages
        are always displayed despite the option you choose.

    CACHE - Cache mode
        Specifies if the ROI and the kernels should be cached. The cache mode
        should be a boolean.

    BINS - The total number of bins
        This is the approximate number of bins. The algorithm finds the best
        possible approximation. For example if BINS = 1000, in a 3D color space
        (e.g. RGB) the histogram will contain 10^3 = 1000 bins, in a 2D color
        (e.g. rgb, where only r and g are used) the histogram will contain
        32^2 = 1024 bins.

    EPSILON - epsilon
        The value of epsilon according to the algorithm described by
        Meer et. al. in the stop condition: |y1 - y0| < epsilon

    COLOR_SPACE - The selected color space
        Several color spaces are available:
            RGB - Red, green and blue channel, ranging from 0 to 1
            rg  - Normalized rgb, i.e. omitting the intensity
            HSV - Hue, satuarion and value, using Matlab's rgb2hsv
            HS  - Hue and saturation, i.e. omitting the intensity
            XYZ - XYZ color space
            xy  - Normalized xyz, i.e. omitting the intensity

    VISUAL_OUTPUT - Display real time result frames and histogram
        This should be a boolean

    ROI_FILE - A stored region of interest
        This should be a file name or []

    RESULT_VIDEO = 'result/result.avi';
        This should be a file name or []

    STATS_FILE - A file to store the object locations y0.
        This should be a file name or []

AUTHORS
    F. Huizinga (folkerthuizinga@gmail.com)
    B. Stoeller    (bramstoeller@gmail.com)

--------------------------------------------------------------------------------

TAGGER

NAME
	tagger - Manually tag the position of an object in each frame of a video

SYNOPSIS
    tracker VIDEO_FILE OBJECT_NAME
    tracker VIDEO_FOLDER OBJECT_NAME

DESCRIPTION
    Note that a roi file for the specified video object is required

--------------------------------------------------------------------------------

CLEANUP

NAME
    cleanup - Cleans the cache

SYNOPSIS
    cleanup

DESCRIPTION
    Empties the cache folder.

--------------------------------------------------------------------------------

TESTALL

NAME
   testall - Run several tests on the tracker

SYNOPSIS
    testall

DESCRIPTION
    In the configuration section of the file you can overload all global
    settings from global.m. 

--------------------------------------------------------------------------------

FIGURES

FIGURES
   figures - Create all figures

SYNOPSIS
    figures

DESCRIPTION
    Creates several graphs and extracts some frames from the result videos.
    This function asumes that ground truth data, statistics and result videos
    are available.