| 摘要 |
PURPOSE:To enable accurate predictive conversion by increasing the number of states even for upper bits relating to the data compression system for compressing data such as character codes and picture data, etc., by arithmetic encoding. CONSTITUTION:A modeling conversion means 10 predetermines distance between adjacent bits inside the same byte and the distance from the same bit position of the byte immediately before for the respective bits inside inputted byte data, the prescribed number of referring bits are selected in the order of the short distance from the same byte and the byte immediately before and the address A of a table pair is obtained from the distance r1-r5 from the bit under consideration to be predictively converted and encoded by an arithmetic encoding means 11. The distance between the referring bit and the bit under consideration can be obtained as Euclidean distance and the distance between the bit under consideration and the adjacent higher rank referring bit, for example, is defined as a unit distance '1'. |
