How to use CImg functions with pixel data?

manpreet Best Answer 2 years ago

I am using Visual Studio and looking to find a useful image processing library that will take care of basic image processing functions such as rotation so that I don't have to keep coding them manually. I came across CImg and it supports this, as well as many other useful functions, along with interpolation.

However, all the examples I've seen show CImg being used by loading and using full images. I want to work with pixel data. So my loops are the typical:

for (x=0;x<width; x++)
for (y=0;y<height; y++)

I want to perform bilinear or bicubic rotation in this instance and I see CImg supports this. It provides a rotate() and get_rotate function, among others.

I can't find any examples online that show how to use this with pixel data. Ideally, I could simply pass it the pixel color, x, y, and interpolation method, and have it return the result.

Could anyone provide any helpful suggestions? If CImg is not the right library for this type of this, could anyone recommend a simple, light-weight, easy-to-use one?

manpreet 2 years ago

You can copy pixel data to CImg class using iterators, and copy it back when you are done.

std::vector pixels_src, pixels_dst;
size_t width, height, n_colors;

// Copy from pixel data
cimg_library::CImg image(width, height, 1, n_colors);
std::copy(pixels_src.begin(), pixels_src.end(), image.begin());

// Do image processing

// Copy to pixel data
pixels_dst.resize(width * height * n_colors);
std::copy(image.begin(), image.end(), pixels_dst.begin());