Glen G. Langdon, Jr. and Jorma J. Rissanen, \A Double Adaptive File Compression Algorithm", IEEE Trans. on Communications, Vol COM-31, No 11, 1253{ 1255, Nov 1983.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....a PL I program that embodies the patent claims. Another description of the Terse algorithm appears in [MW84] The Terse algorithm is the Ziv Lempel (Z L) algorithm used for CE 263 Winter 1998 86 Lempel Ziv Coding Algorithms comparison to the Double Adaptive File Compression (DAFC) algorithm in [LR83b]. An LZW class coder can be optionally used for the type of compression currently used in modems. The GIF algorithm, popularized by CompuServe(R) also uses LZW for the compression of palletized 8 bit index table for color images on the IBM PC class of computer. Because of royalties, a freeware ....

Glen G. Langdon, Jr. and Jorma J. Rissanen, \A Double Adaptive File Compression Algorithm", IEEE Trans. on Communications, Vol COM-31, No 11, 1253{ 1255, Nov 1983.

....We consider strategies that use less space and execute faster but that still achieve better compression than the state of the art compress. Previous Work on Context Models with Modest Memory Requirements Langdon and Rissanen describe a algorithm (LR) that uses a subset of the order 1 model [LR83]. Algorithm LR uses a model consisting of z order 1 contexts and an order 0 context (z is a parameter associated with the algorithm and determines its memory requirements) When encoding begins, the order 0 model is used since no characters have yet occurred in any order 1 context. In a full ....

Langdon, G. G. and Rissanen, J. J. A double-adaptive file compression algorithm. IEEE Trans. Comm. 31, 11 (Nov., 1983), 1253--1255.

....and algorithms that employ the escape strategy of blending provide a much more practical alternative to LOEMA. 2.2. Algorithm DAFC Langdon and Rissanen s Double Adaptive File Compression algorithm is one of the first methods that employs blending in an adaptive data compression scheme [LR83]. DAFC is a partially blended model consisting of z order 1 contexts and an order 0 context (z is a parameter associated with the algorithm and determines its memory requirements) When encoding begins, the order 0 model is used since no characters have yet occurred in any order 1 context. In a ....

Langdon, G. G. and Rissanen, J. J. A double-adaptive file compression algorithm. IEEE Trans. Comm. 31, 11 (Nov., 1983), 1253--1255.

....on line models are doubly adaptive; both the structure and the parameters change to reflect the properties of the data. The separation of modeling from coding, and the use of doubly adaptive stochastic models is the modern data compression paradigm, 1 pioneered by Glen Langdon and Jorma Rissanen [RL81, LR83, Ris83], and further developed in [Ris86a, MGZ89] By paradigm, I mean an ideal and general blueprint for other solutions to follow, and Rissanen s and Langdon s certainly fits this strict interpretation. As mentioned earlier, this paradigm is fully general: it is widely applicable, and it includes ....

....that there are two separate research communities. The first community, represented in information and coding theory literature, has produced the following constructions: ffl The granddaddy of the suffix tree stochastic modeling algorithms is the Doubly Adaptive File Compression algorithm, or DAFC [LR83]. DAFC introduced the now sacred notions of adapting model structure and frequency data simultaneously, and of separating coding from modeling in on line data compression algorithms. Developed in the days of expensive memory, DAFC starts with a complete order 1 m ary suffix tree, then dynamically ....

G. G. Langdon and J. J. Rissanen. A double-adaptive file compression algorithm. IEEE Transactions on Communications, 31(11):1253--1255, 1983.

....interaction is not considered. Models which exploit relationships between source messages may achieve better compression than predicted by an entropy calculation based only upon symbol probabilities. The use of Markov modeling is considered by Llewellyn and by Langdon and Rissanen [Llewellyn 1987; Langdon and Rissanen 1983]. 9. SUMMARY Data compression is a topic of much importance and many applications. Methods of data compression have been studied for almost four decades. This paper has provided an overview of data compression methods of general utility. The algorithms have been evaluated in terms of the amount ....

Langdon, G. G., and Rissanen, J. J. 1983. A Double-Adaptive File Compression Algorithm. IEEE Trans. Comm. 31, 11 (Nov.), 1253--1255.

....We consider strategies that use less space, execute faster and achieve very close to the same compression performance. There are two earlier approaches to the problem of context modeling with modest memory requirements. Contextmodel based algorithms by Abrahamson ( A89] and Langdon and Rissanen ([LR83]) are described in [LH90] These algorithms use order 1 context only and provide compression comparable to that of compress. Neither of these methods compresses as effectively as algorithms FG and PPMC; nor do they generalize naturally to higher order contexts. In order to define a ....

Langdon, G. G. and Rissanen, J. J. A double-adaptive file compression algorithm. IEEE Trans. Comm. 31, 11 (Nov., 1983), 1253--1255.

....have as much statistical context brought to bear, and [rvvp88] was declared the winner. Since then, others have compared their own ideas against HINT, see [db93] hv91] kr92] rc92] Double adaptation techniques (one pass) In file compression, a technique called double adaptation is described in [lr83]. In this technique, the compression takes place in one pass. In the 17 single pass, the algorithm begins compression in the zero order model state. However, it adaptively adds higher order contexts (first order and a special second order runlength context) as their popularity becomes known. ....

G. Langdon and J. Rissanen, "A Double-adaptive File Compression Algorithm", IEEE Trans. on Communications, vol COMM-31, No. 11, November 1983, 1253-1255.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC