Deciphering a code

Question

I am really a novice in C. I need to understand a portion of code whose parameters I am gonna explain below. Suppose we have an image of sizeI and sizeJ which relates to row and column. We are supposed to do cumulative-probability-based histogram equalization method. Can someone explain what this code means?

int histLevelIndex[256];
int histCumuBefore[256];
int N = totalN / 256; // Average N pixel for each level
for(int color=0; color<3; color++)
{
    memset(histCumuBefore,0,256*sizeof(int));
    memset(histLevelIndex,0,256*sizeof(int));
    memset(histogram,0,256*sizeof(int)); // Initialize it into 0
    for(int i=0; i<SizeI; i++)
    {
        for(int j=0; j<SizeJ; j++)
        {
            histogram[ImageData[i][j][color]] ++;
        }
    }
    histCumuBefore[0] = 0;
    for(int i=1; i<256; i++)
    {
        histCumuBefore[i] = histCumuBefore[i-1] + histogram[i-1];
    }
    // No explicit transfer function used here, since transformation of some pixel is based on pixel's order
    for(int i=0; i<SizeI; i++)
    {
        for(int j=0; j<SizeJ; j++)
        {
            histLevelIndex[ImageData[i][j][color]] ++;
            int L = (histCumuBefore[ImageData[i][j][color]]+histLevelIndex[ImageData[i][j][color]]-1) / N;
            if(L>=256)
                L = 255;
            ImageDataOut[i][j][color] = L;
        }
    }

I know that the first part is doing histogram and cumulative histogram. Can you please explain what the next part is doing?

 for(int i=0; i<SizeI; i++)
    {
        for(int j=0; j<SizeJ; j++)
        {
            histLevelIndex[ImageData[i][j][color]] ++;
            int L = (histCumuBefore[ImageData[i][j][color]]+histLevelIndex[ImageData[i][j][color]]-1) / N;
            if(L>=256)
                L = 255;
            ImageDataOut[i][j][color] = L;
        }
    }

I would be grateful for any sort of help. Thanks.

Since you've tagged this post as MATLAB, I'm assuming you're open to solutions in that language as well. You first need to **understand** how histogram equalization works. Once you do that, the code is very easy to understand. That C code is straight up histogram equalization. Please read the duplicate and figure out how it works first. Once you do that, the code will be simple to understand. — rayryeng, Sep 20 '15 at 03:31
I know how transfer-function based equalization works. But I want to manipulate the values using bucket-filling method, which basically means that I want a histogram of equal length for all pixels. Do you know how that works? — Nasiba, Sep 20 '15 at 03:37
Here is the thing, I don't understand the intuition of how that works on coding? — Nasiba, Sep 20 '15 at 03:46

Deciphering a code

0 Answers0