Comparison of Huffman and Arithmetic Coding in Image Compression

Cilj rada je usporedba dva najpoznatija algoritma za entropijsko kodiranje, Huffmanovog i aritmetičkog. Točnije, usporedba je temeljena na tome koji od ta dva algoritma je bolji kod kompresije slike u JPEG formatu. Sukladno tome, ukratko je opisan način rada JPEG-a te uloga Huffmanovog i aritmetičkog algoritma u tom postupku. Huffmanov i aritmetički algoritam su opisani tekstualno, dijagramom toka, jednostavnim primjerom i izvedbom u programskom jeziku Python, a zatim su primijenjeni nad konkretnim podacima koji nastaju pri kompresiji slike neposredno prije entropijskog kodiranja. Parametri prema kojima se provodila usporedba jesu stupanj kompresije i složenost algoritama. Na temelju tih podataka i rezultata dobivenih testiranjem algoritama, zaključeno je da je aritmetički algoritam pogodniji u postupku kompresije slike od Huffmanovog algoritma.The object of this paper is to compare the two most known algorithms for entropy coding, Huffman and arithmetic. In more particular, comparison is obtained between these two algorithms in order to prove which one is more apropriate for image compression in JPEG format. Accordingly, the JPEG and the part of Huffman and arithmetic algorithm in its process are briefly described. Descriptions of algorithms are performed by using text, flow chart, simple example and execution in the Python programming language, and then applied to the real life data generated by the step which leads entropy coding in image compression. The comparative parameters of algorithms are the compression ratio and the complexity of the algorithms. Based on these data and testing results, conclusion is made that arithmetic algorithm is more useful than Huffman algorithm in the image compression process