Convert the integer vector I into its binary matrix representation (8 bits per value, most-significant bit first). Return the (len(I), 8) array of 0s and 1s.