Context-Adaptive Binary Arithmetic Coding With Fixed-Length Codewords

Abstract

Context-adaptive binary arithmetic coding is a widespread technique in the field of image and video coding. Most state-of-the-art arithmetic coders produce a (long) codeword of a priori unknown length. Its generation requires a renormalization procedure to permit progressive processing. This paper introduces two arithmetic coders that produce multiple codewords of fixed length. Contrary to the traditional approach, the generation of fixed-length codewords does not require renormalization since the whole interval arithmetic is stored in the coder's internal registers. The proposed coders employ a new context-adaptive mechanism based on variable-size sliding window that estimates with high precision the probability of the symbols coded. Their integration in coding systems is straightforward as demonstrated within the framework of JPEG2000. Experimental tests indicate that the proposed coders are computationally simpler than the MQ coder of JPEG2000 and the M coder of HEVC while achieving superior coding efficiency.Index Terms-Context-adaptive binary arithmetic coding, fixedlength arithmetic codes.

GPU Implementation of Bitplane Coding with Parallel Coefficient Processing for High Performance Image Compression

“…Arithmetic coding is an entropy coding technique extensively employed in the coding field due to its high efficiency [23]. From an algorithmic point of view, an arithmetic coder divides an arithmetic interval in two subintervals whose sizes are proportional to the estimated probability of the coded bit.…”

“…The main difference is that it generates multiple fixed-length codewords instead of a single and long codeword that has to be processed in small segments [23]. The fixedlength codeword arithmetic coder is adopted by BPC-PaCo because it reduces computational complexity and control flowdivergence.…”

Bitplane Image Coding With Parallel Coefficient Processing

“…The coefficients coded by a thread are visited in a sequential order, so an arithmetic coder can be individually employed to code all symbols emitted for a stripe. Instead of using conventional arithmetic coding, we employ an arithmetic coder that generates codewords of fixed length [39]- [43]. Variable-to-fixed length arithmetic coding avoids renormalization, reducing the complexity of the coder [43].…”

“…Instead of using conventional arithmetic coding, we employ an arithmetic coder that generates codewords of fixed length [39]- [43]. Variable-to-fixed length arithmetic coding avoids renormalization, reducing the complexity of the coder [43]. It uses an integer interval with a pre-defined range, say [0, 2 W − 1] with W being the length of the codeword (in bits).…”

“…In this procedure, I[t] is the codeword read from the bitstream for stripe t. The procedure is similar to that of the encoder. An extended description of the arithmetic coder employed in Algorithms 1 and 2 can be found in [43].…”