Method and apparatus for compressing digital data using non-adaptive predictive techniques

An apparatus for the processing of digital data from an image, comprising:

• means for converting said digital data into varying run lengths of a first type data element and a second type data element;
• first receiving means for receiving from said conversion means a serial stream of said first and second types of data elements, including first storage means for storing a specified number of data elements in a pre-defined pattern at any particular time;
• second storage means, coupled to said first storage means, for storing a plurality of predicted data elements which are outputted in response to a first address corresponding to a specific arrangement of data elements within said first storage means;
• third storage means, coupled to said first storage means, for storing a plurality of source state elements, each of which comprises one of a set of source states, which are outputted in response to said first address corresponding to said specific arrangement of data elements within said first storage means, and which are indicative of the degree of probability that a predicted data element is in error;
• first comparison means, coupled to said conversion means and said second storage means, for comparing each of said predicted data elements from said second storage means to its corresponding original data element from said conversion means to generate a plurality of prediction error data elements, each of said prediction error data elements maintaining the same relationship to a given source state element as its corresponding predicted data element;
• first control means, coupled to said first comparison means, and responsive to each prediction error data element and its corresponding source state, for generating a set of first signals;
• first counter means, responsive to said set of first signals, for counting said prediction error data elements received by said first control means according to type and source state;
• memory means, coupled to said first counter means, for arranging said prediction error data elements into a plurality of groups in each of said source states, said groups constituting digital symbols representative of each group;
• first addressing means for producing an address to said memory means in order that said symbols may be stored in said memory means according to the sequential order of occurrence of said source states;
• fourth storage means coupled to said memory means, for storing a plurality of code words which are representative of each of said symbols;
• second control means, coupled to said fourth storage means, for causing the outputting of selected code words from said fourth storage means;
• and means for outputting said selected code words from said fourth storage means over a transmission medium.

A method for processing digital data from an image, comprising the steps of:

• (a) converting said digital data into varying run lengths of a first type data element and a second type data element;
• (b) receiving a continuous stream of said digital data and storing said data in a first storage means, which first storage means stores a specific number of data elements in a predefined pattern at any particular time;
• (c) generating a predicted data element and a corresponding source state element, which source state element represents one of a set of source states for each original data element received from said image, where said predicted data element reflects the probable type of data element from the image with respect to a specific arrangement of other data elements within said storage means, and said source state element reflects the degree of probability that said corresponding predicted data element is in error;
• (d) comparing each predicted data element to its corresponding original data element from said image to generate a prediction error data element, where said prediction error data element is associated with the same source state as its corresponding predicted data element;
• (e) counting said prediction error data elements and storing said elements according to type and source state in order to form a plurality of symbols in each of said source states;
• (f) inputting each of said symbols to a storage device as an address thereof;
• (g) outputting from said storage device code words representative of each of said digital symbols;
• (h) and transmitting said code words over a transmission medium.