Hi C4R0,
Hahahah, man you rock, i just downloaded it, just to see the HOBBYBOB button !
Since i just got home from a party and it's already 3.39am here i will check tomorrow.
Nice , thanks !
Hi C4R0,
Hahahah, man you rock, i just downloaded it, just to see the HOBBYBOB button !
Since i just got home from a party and it's already 3.39am here i will check tomorrow.
Nice , thanks !
I didn't fail !
I just found out 10,000 ways that didn't work.
HiI made some upgrades to my little program, mostly improovments of the interface. Here's the link: ftp://plftp:ildaswap@ftp.photonlexic...bmp2ild_v2.exe
And some sample files: ftp://plftp:ildaswap@ftp.photonlexic...mple_files.zip
Since there's nothing new in "the hobbybob button" function and it still works crappy, I disabled it. I hope to make this function work better and add multiframed ILD files support in the next version.
http://c4r0.byethost9.com/ - If you like lasers, high voltage, x-rays and stuff like that, check out my website
could you add other formats to it not just bmp, like jpg, gif and png these are the most common once.
Thats a very clean conversion CR40. The other programmes I've seen that trace don't seem to produce anything like as accurate a result. Only comment I would make is that it seems a little point heavy on the straight lines although not doubt a post save optimiser such as ILDA SOS would soon sort that out.
This looks awsome. I can't wait to give it a try.
I'm not even sure what you're talking about :P I don't recognize vector paths. Let's take a white shape on black background for example. When you click on the background, the program looks for the closest white pixel. When it's found, the program checks eight pixels around one by one. When it detects an edge, which is simply a change from black to white, it goes to that pixel and start checking eight pixels around again. That procedure is repeated until it reaches the pixel it started from. It's a simple algorithm written by me - a guy who knows nothing about edge detection or tracking :P
Looks like the background is not actually blackIt has to be really black (that means RGB 0,0,0 ), every other colour is recognized as an object. I'm planning to add some thresholding to color checking procedure but it's not made yet.
http://c4r0.byethost9.com/ - If you like lasers, high voltage, x-rays and stuff like that, check out my website
Well, it is the simplest form of wave algorithm(possibly it has another name in English?). Many developers use it in their apps which are intended to seek paths(game developers etc.). I use recursive function calls. It requires a lot of stack space but simple to realize.
Why do you use scanning only in one way? I use vertical+horizontal lines scanning and it works much better.
I assume it's seeking for colored shapes in the picture, not concerned with tracking edge after a shape is found, right? Then, when you use this method, how do you know when to stop scanning? When using simply line by line scanning I know when I checked all pixels, and thus when to stop. In case of scanning by lines from one pixel in eight directions like in the picture, you can easily ommit some small shape that happens to fit between the lines. Then, when all eight lines reaches image borders without discovering any new shape, you think that you found all, but it's wrong. ... Do I understand it at all?![]()
http://c4r0.byethost9.com/ - If you like lasers, high voltage, x-rays and stuff like that, check out my website
I don't try to found any shape. Just do some frequency filtration to get outlines. After frequency filtration I have 2 types of points(in bitmap) with different status: points which are outlines and black points(barrier). The tracer can move only between outline points. When it traces the next point then it is marked as barrier point and it's coordinates are placed in stack. It prevents repeated tracing. After tracing last point the curve locks and tracer start to find new curve with non-traced points. Thus I have to look trough H x W pixels, where H=height and W=width of image.
At the end of this stage I have points one-by-one stored in stack with step of 1 pixel. Then I apply point reducing procedure. However it is far from perfect and points' density more than makes your program.
Similar to your app but another implementation.