[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## [Help-gsl] lssolve precision

**From**: |
potda |

**Subject**: |
[Help-gsl] lssolve precision |

**Date**: |
Mon, 25 May 2009 11:04:17 +0200 |

Hello! I'm writing my final thesis about Image compressing. In the
algorithm I use the gsl_linalg_QR_lssolve routine to find whether a quadrant of
the image can be linearly expressed by a basis.The number of rows in the matrix
A ( from the A*x=b problem) is greater or equal to the number of
columns.Otherwise we can not assume nothing about it except that the values are
between 0 and 255 and are integers. The problem is that the precision of the
solution is not satisfying therefore on images greater than 128x128 pixels
appears some unwanted pixels. It looks like the greater is the basis ( which
means that A has more cols, more than 100) the more errors i find. Can I
achieve a more precise solution using the gsl library? Which method should I
use? Maybe I should somehow manipulate matrix A (normalize or whatever)?
Thank You! Daniel
--------------------------Hirdetés-----------------------------
Mostantól könnyebb átköltözni az Indamailbe!
Az ismerőseid listáját pár kattintással áthozhatod.
2 GB tárhely, testreszabható működés és megjelenés.
Indamail.

**[Help-gsl] lssolve precision**,
*potda* **<=**