Steve on Image Processing

April 23rd, 2010

More R2010a performance improvements in Image Processing Toolbox

Earlier this week I posted some relative speed improvement charts for some functions in the Image Processing Toolbox that got faster in R2010a.

Below are some more charts. I've included applylut, bwulterode, ifanbeam, imclearborder, imdilate, imerode, medfilt2, rangefilt, and stdfilt.

There are other functions that have gotten faster in the last couple of releases, including bwpack, bwselect, cornermetric, decorrstretch, edge, imclose, imopen, imextendedmax, imfill, imhmax, imimposemin, improfile, imreconstruct, imrotate, and imtransform.

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

performance improvement chart

3 Responses to “More R2010a performance improvements in Image Processing Toolbox”

  1. Royi replied on :

    Those are great results.
    Yet, the dream is running those functions on the GPU.

  2. Petter replied on :

    I agree, Mathworks should buy Accelereyes and incorporate Jacket as a standard toolbox. That would make MATLAB much more competitive. GPU computing is here to stay.

  3. Andrew Scott replied on :

    I also agree.

    I have, however, heard rumours that we should see something in a not too distant release.


MathWorks
Steve Eddins is a software development manager in the MATLAB and image processing areas at MathWorks. Steve coauthored Digital Image Processing Using MATLAB. He writes here about image processing concepts, algorithm implementations, and MATLAB.

These postings are the author's and don't necessarily represent the opinions of The MathWorks.