Suggestions for perf improvement on small images?

[dxclancy]

Hi, I was just wondering if there was something relatively simple I could do to increase performance when using small images as input.

I'm not saying that ColorCube isn't performant. But i am calling it many times, and was just wondering if perhaps for small images there is something obvious. Like "oh, you really don't need to do x passes, or this data structure could be smaller, or if your images are that small just do it yourself like this"

Essentially, I am almost exclusively working with 10x10 images and looking for only the first dominant color.

thanks!!

[Ole]

Hi, image size is the main factor when it comes to speed. So you are doing that right :)

You should make sure that you create the color cube instance once, and reuse that. This way data structures do not get re-allocated. So do this only once:

CCColorCube *colorCube = [[CCColorCube alloc] init];

Otherwise there is not much you could do about it. It does not really make a different if you are looking for only the first dominant color or more. In both cases the whole image must be processed in the same way.

[Ole]

Does that help?

[dxclancy]

Not really :) but I appreciate the answer!! Thanks!