Hi Harry,
While Blitz++ is not specific to image processing, the
predecessor to Blitz++ WAS an image processing library,
so you may find many useful features. For example,
you can calculate the L2-distance between two images using
this bit of code:
Array<float,2> A(N,N), B(N,N); // two images
float distanceL2 = sum(pow2(A-B));
Or the L1 distance:
float distanceL1 = sum(abs(A-B));
Cheers,
Todd
> I am working with data sets which consist of small greyscale
> images, typically 48 x 48 pixels on the small side and
> seldom larger than 100 x 100 pixels. This data comes from
> electron micrographs and the pixel values are stored and worked
> with as floats. Typically there are more than 500 images,
> all with the same dimensions in pixels, in a dataset that I
> work with at any time. For example, I could have 1000 images,
> each image 64 x 64 pixels, that I need to classify using some
> clustering algorithm.
>
>
> I am currently rewriting some of the work from my thesis and
> rather than use the class I developed for images in my thesis
> I was going to write a new one. I was thinking of using blitz++
> since it would allow me to use a two dimensional array as an
> aggregate member to store the data of a single image.
>
> Has anyone tried to use blitz to work with greyscale images?
>
> Any suggestions?
>
> Harry Zuzan
>
> --------------------- blitz-support list --------------------------------
> * To subscribe/unsubscribe: mail to majordomo@oonumerics.org, with
> "subscribe blitz-support" or "unsubscribe blitz-support" in the body of the message
> * Blitz++ web page: http://oonumerics.org/blitz/
>
>
--------------------- blitz-support list --------------------------------
* To subscribe/unsubscribe: mail to majordomo@oonumerics.org, with
"subscribe blitz-support" or "unsubscribe blitz-support" in the body of the message
* Blitz++ web page: http://oonumerics.org/blitz/
This archive was generated by hypermail 2b29 : Wed Feb 20 2002 - 05:10:05 EST