Content deleted Content added
Line 147:
ok, touche, n i think u r not such well intended regarding my work, somehow i think i know people like u, here in Romania: please watch ur style talking to me, ok:just a friendly advise, n i mean it !!!! [[Special:Contributions/93.118.212.93|93.118.212.93]] ([[User talk:93.118.212.93|talk]]) 13:12, 7 April 2013 (UTC)
now B a sport n tell to the people witch weel is those :
hi, i think that this old O(1) floodfill algo that trades time 4 mem, might b help in time consuming by backtracking concepts of smart test, backjumping so where is no necessary to walk on the same line bcz we already know by pasts walks there is no reason, we can simply add a marker to help comparing forward next line of forward bacvtracking algo with the experience from last pass... we might evn reach O(N) time, if we r a bit lucky 93.118.212.93 (talk) 11:43, 19 January 2013 (UTC)
hi, i think there could be written algos based on grid concept for the stored values, n a small square 4 data, also, that possibly work in O(N) time n O(N^(2/3)) mem, for example or even less O(N^(1/2)), O((log(N)), or evn O(1) mem for O(N) time, so i dont think really its need to trade time for mem . for example O(1) mem algo first seek for target without filling then intelligent border keeping also condition of conexivitty (local test) 4 the remaining zone... only if there is no bordering n need to spilt, then proceed splitting... O(N^1/2) of O(N^2/3) are evn more simple they use O(N) time. it some cautious terms N is the maximum number of pixel of a possible image to b filled... — Preceding unsigned comment added by 93.118.212.93 (talk) 10:29, 18 January 2013 (UTC)
| |||