use a convolution kernel size based on the stddev rather than a
authorTony Cook <tony@develop=help.com>
Mon, 9 Apr 2007 13:55:11 +0000 (13:55 +0000)
committerTony Cook <tony@develop=help.com>
Mon, 9 Apr 2007 13:55:11 +0000 (13:55 +0000)
commit73962964481c4e85fc999d7944fee872cc391238
treea5847e44414296f432b475d457d72dfc6669cb64
parentf2ad77de22ae8f74276412c2e6bb45acb55cdbeb
use a convolution kernel size based on the stddev rather than a
fixed size when performing a gaussian blur

also uses the i_fcolor interfaces when working with >8 bit/sample
images
RT 25645
Changes
MANIFEST
gaussian.c [deleted file]
gaussian.im [new file with mode: 0644]
t/t61filters.t