Dr. Jennifer Burg

Department of Computer Science

Research Project for Undergraduates

Entropy Encoding for Digital Image Compression

Entropy encoding is a compression method based on the idea of allotting fewer bits to represent colors that occur frequently in an image and more bits to those that occur infrequently. Shannon’s entropy equation, allows us to compare an encoding to a theoretical optimum. This research project involves implementing entropy encoding algorithms such as Shannon-Fano, Huffman, and arithmetic encoding and comparing their performance on different types of images. Arithmetic encoding, which theoretically can encode an entire image in one real number, is of particular interest. This appears to require infinite precision for the real numbers. How arithmetic encoding is made feasible will be studied. For more information, see http://www.cs.wfu.edu/%7Eburg/nsf-due-0340969/textbased/Chapter1CS.pdf.