Back to results
US 5,708,431 AGranted
Method for compression coding of potentially unbounded integers
AssigneeLucent Technologies Inc.
InventorPaul Glor Howard
Life of this patent
Estimating the term…
Loading patent…
Loading patent…
AssigneeLucent Technologies Inc.
InventorPaul Glor Howard
Estimating the term…
An integer of potentially unbounded size is compression coded by first transforming the integer into a binary representation in which the length of the sequence of bits representing the integer is self-contained within the sequence itself. Each of the bits within the sequence is then coded with a binary arithmetic coder, such as the QM-Coder, which uses as a conditioning context for each bit, the bits of the sequence previously coded. In order to limit the amount of memory used to store the probability information associated with each conditioning context, memory is allocated for storing such information as that conditioning context is needed. Also, once a predetermined amount of memory has been used to store this probability information for different conditioning contexts, no further memory space is allocated for newly occurring conditioning contexts. Rather a special overflow memory space is allocated and used for all conditioning contexts not previously defined.
A method of compression coding an integer of potentially unbounded magnitude comprising the steps of:
A method of compression coding a non-negative integer or a negative integer n of potentially unbounded value comprising the steps of:
Dependencies are inferred from the claim text.