An Ordering Scheme for Facsimile Coding

01 November 1979

New Image

This paper presents simulation results using one of our ordering schemes 1,2 for efficient coding of two-level (black-and-white) facsimile pictures using the eight CCITT (International Telegraph and Telephone Consultative Committee) test documents* as the source data. The scheme, we consider, allows exact reproduction of the documents at the receiver. Specifically, we give results on compression factors using a seven-element predictor, and a modified Huffman code for coding of the ordered data. It is assumed that every kth. (k = 2, 4, o°) line is encoded by a one-dimensional run-length code to limit the vertical propagation of the transmission errors. In our previous papers, 1 3 we gave entropies of run lengths of the ordered data using predictors and * The eight C C I T T study group xiv test documents are those used for the Graphics Coding Contest of the 1976 Picture Coding Symposium. Each image contains 2128 lines with 1728 pels per line. The scanning density in both directions is approximately 8 dots/mm. 2113