From: D M German <dmg@uv...> - 2009-08-14 18:38:43
I just committed a patch that fixes the g-and-t parameters.
* Updated tests cases to reflect the slight change in the
boundaries of the ROI.
* adjust.c, math.c, filter.h: The inverse of shear was broken. I
have added a shearInv function that takes care of this bug.
* PTcommon.c (getROI): Improved its computation of its edges to
make it err on the outside rather than on the inside of the actual
area. I have also added code to test the inverse computations for
any particular function.
I have also added extra assertions to getROI to make sure that the
inverses are working properly.
Whenever somebody releases a final version, please compile it without
NDEBUG (although the extra computations are not really that expensive).
Please let me know if anybody finds a problem. I suspect most people do
not use PTmender these days.
Daniel M. German
dmg (at) uvic (dot) ca
replace (at) with @ and (dot) with .
Get latest updates about Open Source Projects, Conferences and News.