kd-forest.git - An implementation of k-d forests

	Commit message (Collapse)	Author	Age	Files	Lines
...
*	Make kdf entirely internal to generate_image.	Tavian Barnes	2014-05-11	1	-20/+21
\|
*	Refactor main.c.	Tavian Barnes	2014-05-11	1	-88/+128
\|
*	Allow all possible bit-depths, not just multiples of three.	Tavian Barnes	2014-05-10	1	-9/+12
\|
*	Use only integer math for hue comparisons.	Tavian Barnes	2014-03-20	1	-18/+18
\| \| \| \| \|	This is more than 10x faster than how it started, and over 5x faster than the previous commit.
*	Fix corner case in color_comparator() and clarify a comment.	Tavian Barnes	2014-03-20	1	-4/+4
\| \| \| \| \| \| \|	If anum == 0 and bdenom == 0, we should check anumsgn(adenom) < bnumsgn(bdenom), not anumadenom < bnumbdenom.
*	Use link-time optimization.	Tavian Barnes	2014-03-20	1	-1/+1
\|
*	Make hue sorting 3x faster by avoiding atan2() calls.	Tavian Barnes	2014-03-20	1	-16/+51
\|
*	Put -lm -lpng at the end of the command line.	Tavian Barnes	2014-03-13	1	-3/+6
\|
*	Be more const friendly.	Tavian Barnes	2014-03-12	2	-5/+5
\|
*	Try to be more even in which neighbor to choose.	Tavian Barnes	2014-03-12	1	-18/+23
\|
*	Look for strictly better matches in the k-d tree code.	Tavian Barnes	2014-03-11	1	-2/+2
\|
*	Refactor hue comparator into color.c.	Tavian Barnes	2014-03-11	3	-28/+27
\|
*	Get rid of some unreachable abort()s.	Tavian Barnes	2014-03-11	1	-6/+0
\|
*	Allow different color spaces to be used for similarity measurement.	Tavian Barnes	2014-03-11	6	-64/+216
\|
*	Use a better method of getting rid of HUE_SORT artifacts.	Tavian Barnes	2014-03-11	1	-9/+6
\|
*	Initial commit.	Tavian Barnes	2014-03-11	8	-0/+734