What to do with the output of the DCT step in JPEG image compression?

I am working on JPEG algorithm. After the dct step, the class is converted to double from unit8. In the zigzag step, the matrix is converted to an one dimensional array. I am confused now as how to get back the unit8 image class again after the compression? What does the huffman coding part do? I am working on black and white jpeg image. But what if I take a color image as after dct, I guess it turns to black and white.