Alright, I got my program working with plain delta iteration. It's able to recognize glitches, however it's not correcting them yet. It currently uses the center of the screen as a reference point. This is not optimal, but I don't really know what to do here. How do I find a relatively good reference point to start with (given the position in the complex plane and a zoom factor)? After the first pass is rendered, do I just use the center of the glitch blobs as new reference point or is there a better way to do this? I don't really understand how SFT is doing it.
Here's what I got so far (red means glitch).
...
The center of a blob is not the best point, you need the locus where the highest iterations are, in the simplest terms one might proceed as follows...
Do a very loose scan, say two or three horizontal and vertical lines through the data rendered from the first pass looking for large blobs
if found, estimate locus and use as reference point, render pixels marked as glitch, repeat
if no more big blobs...
Do a more dense scan looking for smaller blobs
if found, estimate locus and use as reference point, render pixels marked as glitch, repeat
if no more small blobs...
Do a scan of remaining pixels and use the highest iteration location as a reference, render pixels marked as glitch
by this time you should be done
Is K.I.Martin's SFT solving glitches at all?
When I ported SFT(java) to C++ I found that the reference point selection wasn't working as well as I though it should and I couldn't quite figure it out so I scrapped the original java version in favor of my own "hunt an peck" method that finally seems to work as of v0.75
@nitroxis
if want to look at the souce 
June 02, 2009, 09:58:38 PM
Dilber MC Theme by HarzeM