Huffman Coding Method Based on Number Character

This paper presents a new Huffman coding method based on number character. The traditional 256 code table is replaced by the 0-9 character, the space character and the enter character in this method. Meanwhile it illustrates that the value of entropy has close relations with the probability model of signal; that the same signal has various entropies under different models; and that the probability model of signal which makes the entropy smaller can provide bigger compression space. Moreover this approach not only reduces the size of traditional Huffman code table but also enhances the compression ratio of the image data to some extent. A large number of experimental data from random online traffic images indicate that absolutely the Huffman code table can be controlled below 40bit, while the coding efficiency can be upwards of 95% and the compression ratio will exceed 60%.