Fast 2D histogram calculation

Fast calculation of 2D histogram. Optionally displays surf plot of histogram.

Description

Hist2D performs a fast calculation of 2D histograms. On a 2GHz processor, calculated and displayed a histogram of 1 million points in less than half a second.

HIST2D 2D histogram
 
  [H XBINS YBINS] = HIST2D(D, XN, YN, [XLO XHI], [YLO YHI])
  [H XBINS YBINS] = HIST2D(D, 'display' ...)
 
HIST2D calculates a 2-dimensional histogram and returns the histogram array and (optionally) the bins used to calculate the histogram.
 
Inputs:
      D: N x 2 real array containing N data points or N x 1 array of N complex values
      XN: number of bins in the x dimension (defaults to 20)
      YN: number of bins in the y dimension (defaults to 20)
      [XLO XHI]: range for the bins in the x dimension (defaults to the minimum and maximum of the data points)
      [YLO YHI]: range for the bins in the y dimension (defaults to the minimum and maximum of the data points)
      'display': displays the 2D histogram as a surf plot in the current axes
 
Outputs:
      H: 2D histogram array (rows represent X, columns represent Y)
      XBINS: the X bin edges (see below)
      YBINS: the Y bin edges (see below)
        
As with histc, h(i,j) is the number of data points (dx,dy) where x(i) <= dx < x(i+1) and y(j) <= dx < y(j+1). The last x bin counts values where dx exactly equals the last x bin value, and the last y bin counts values where dy exactly equals the last y bin value.
 
If D is a complex array, HIST2D splits the complex numbers into real (x) and imaginary (y) components.

Acknowledgements

The author wishes to acknowledge the following in the creation of this submission:
2D Histogram

• MATLAB release: MATLAB 7.7 (R2008b)
• Other Files: hist2d.m, ; license.txt

[Bruno Luong]

Speed is decent, but still cannot beat another more generic histogram submission:

>> tic; [Hout xbins ybins] = hist2d(x); toc
Elapsed time is 0.282145 seconds.

>> tic; [count xbins ybins] = histcn(x,xbins,ybins); toc
Elapsed time is 0.218104 seconds.

http://www.mathworks.com/matlabcentral/fileexchange/23897-n-dimensional-histogram

Let alone an even faster mex implementation
http://www.mathworks.com/matlabcentral/fileexchange/3957

So I would change the implicit speed claim in the title.

Bruno