Histgram Equalization Algorithm Notes

w:        width of image

h:        height of image

Lmin :  Min gray level of some gray image, relevant to image bit depth, in generally, 0

Lmax: Max gray level of some gray image, relevant to image bit depth, in generally, 255

 

1. statistics occuring frequency of gray level

    for(i = 0; i < h; i++)

          for(j = 0; j < w; j++) n[graylevel[i][j]]++;

   

2. normalization

    for(i = 0; i < Lmax; i++) p[i] = n[i] / (w * h);

 

3. accumulate probability

    p[i] += p[i - 1];     i = 0, ..., Lmax

 

4. equalized image

    for(i = 0; i < h; i++)

         for(j = 0; j < w; j++) newgraylevel[i][j] = p[graylevel[i][j] * (Lmax - Lmin) + Lmin