class TFBaseImage: public TFIOElement

 Default constructor. The image has 0 dimension and is not
 associated with an image in a file.
 Don't use this constructor. The image should have at least a name.

 copy constructor

 Creates a 1 - dimensional image in memory

 Creates a 2 - dimensional image in memory
 dim2 is the most frequently changing dimension, i.e X
 to access one pixel use image[dim1][dim2]

 Creates a 3 - dimensional image in memory
 dim3 is the most frequently changing dimension, i.e X
 to access one pixel use image[dim1][dim2][dim3]

 Creates a n - dimensional image in memory
 to create a 5 - dimensional image one can write:
    UInt_t size[5] = {4, 6, 3, 10, 15};
    TIntImage * img = new TIntImage("name", 5, size);
 Again the most right index (in the example the value 15) is the most
 frequently changing dimension

 Creates a new 1 - dimensional image in memory and associates it with a new
 image in a file.

 Creates a new 2 - dimensional image in memory and associates it with a new
 image in a file.
 dim2 is the most frequently changing dimension, i.e X
 to access one pixel use image[dim1][dim2]

 Creates a new 3 - dimensional image in memory and associates it with a new
 image in a file.
 dim3 is the most frequently changing dimension, i.e X
 to access one pixel use image[dim1][dim2][dim3]

 Creates a new 3 - dimensional image in memory and associates it with a new
 image in a file.
 to create a 5 - dimensional image one can write:
    UInt_t size[5] = {4, 6, 3, 10, 15};
    TIntImage * img = new TIntImage("name", 5, size);
 Again the most right index (in the example the value 15) is the most
 frequently changing dimension

 destructor

 protected funcion to initialze some arrays.

 Returns the size in each dimension. The array of size must be large enough
 to hold a value for each dimension. Use GetNumDim() to ask for the number
 of dimesnions of the image.
 If sub == kFALSE (the default) the size of the original image is returnd
 independent of a previously call of the MakeSubSection() - function.
 If sub == kTRUE the size of a sub - image is returned which was previously
 defined with the MakeSubSection() - function.
 In any case size[0] will hold the size of the least frequently changing
 dimension. For example for a 2 - dimensional image the Y - axis.

 returns the total number of pixels of the original image (sub == kTRUE)
 or of a previously defined sub - image (sub = kTRUE)

 Defines a subsection of an image.
 begin defines the first pixel in each dimension of the original
 image which will be become the index 0 of the sub section.
 end defines the pixel in each dimension of the original
 image which will be behind the last pixel of the sub - image.
 The size of the sub - image in a dimension will be
 end[dimX] - begin[dimX]
 The first index ( index 0 ) of begin and end defines the least
 frequently changing dimension. For example in a 2 - dimensional
 image the Y - axis.

 A dimension can be freezed if begin[dimX] == end[dimX]
 A pixel of the sub - image must be accessed with the operator ()
 instead of the opertor []. Frozen dimensions must be skipped.

 For example:
    TFIntImg img("name", 10, 5, 20);
    UInt_t begin[3] = {3 , 3, 10};
    UInt_t end[3]   = {10, 3, 15};
    img.MakeSubSection(begin, end);

 The sub - image of img has 2 dimensions and is of size  7 X 5
 img[4][3][12] will now access the same pixel as img(1)(2)

 As second call of this function will first reset the image to its
 original size and than apply the begin and end - values.

 Resets a previously resized image ( fucntion MakeSubSection() ) to its
 original size.

 Create a new histogram of the required type
 Only one and two dimensional histograms are supported
 If a sub - image is defined ( function MakeSubSection() ) the histogram
 will be created using only the sub -image, else the full image is
 used. If the image has more dimensins than the histogram the least
 frequently dimensions of the image are used.
 For example to create a two dimensional histogram of the X - and Z - dimension
 from a 3 - dimensional image first a sub - section with a freezing
 third dimension has to be defined. For example:

    UInt_t begin[3] = {3,  2, 20};
    UInt_t end[3]   = {35, 2, 432};
    img.MakeSubSection(begin, end);
    TH2D * hist = dynamic_cast<TH2D*>(img.MakeHisto());

 The parameter type defines the Class type of the returned histogram.
 For example:
    TH2F * hist = (TH2F*)MakeHisto( TH2F::Class() )

 The returned histogram has to be deleted by the calling function

 Create a new histogram of the required type
 Only one and two dimensional histograms are supported

 If this image is has 3 dimensions (z,y,x) and the histogram has 2
 dimensions (y,x) the parameter zPos defines the z - position which
 is used to build the histogram. zPos = 0 defines the first layer.
 In any other case the value of zPos has not affect. To build a
 subcube of the images use the function MakeSubSection() as described
 hereafter.

 If a sub - image is defined ( function MakeSubSection() ) the histogram
 will be created using only the sub -image, else the full image is
 used. If the image has more dimensins than the histogram the least
 frequently dimensions of the image are used.
 For example to create a two dimensional histogram of the X - and Z - dimension
 from a 3 - dimensional image first a sub - section with a freezing
 third dimension has to be defined. For example:

    UInt_t begin[3] = {3,  2, 20};
    UInt_t end[3]   = {35, 2, 432};
    img.MakeSubSection(begin, end);
    TH2D * hist = dynamic_cast<TH2D*>(img.MakeHisto());

 The parameter type defines the Class type of the returned histogram.
 For example:
    TH2F * hist = (TH2F*)MakeHisto( TH2F::Class() )

 The returned histogram has to be deleted by the calling function

 cretes a TTree. One branch is the pixel value the other branches are
 the axis of this image. There is one record per image pixel which are
 not NULL. If a sub - image is defined the tree will be build only of
 the pixels of this sub - image.
 nameConvert can be NULL. But it will be adopted by this
 function and will be deleted by this function if it is not NULL.
 The calling function has to delete the returning tree.