Entropy coding in video compression using probability interval partitioning
Detlev Marpe, Heiko Schwarz, Thomas Wiegand
Abstract
Detlev Marpe, Heiko Schwarz, Thomas Wiegand
Abstract
We present a novel approach to entropy coding, which provides the coding efficiency and simple probability modeling capability of arithmetic coding at the complexity level of Huffman coding. The key element of the proposed approach is a partitioning of the unit interval into a small set of probability intervals. An input sequence of discrete source symbols is mapped to a sequence of binary symbols and each of the binary symbols is assigned to one of the probability intervals. The binary symbols that are assigned to a particular probability interval are coded at a fixed probability using a simple code that maps a variable number of binary symbols to variable length codewords. The probability modeling is decoupled from the actual binary entropy coding. The coding efficiency of the probability interval partitioning entropy (PIPE) coding is comparable to that of arithmetic coding.
OpenAlex reports 20 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
We present a novel approach to entropy coding, which provides the coding efficiency and simple probability modeling capability of arithmetic coding at the complexity level of Huffman coding. The key element of the proposed approach is a partitioning of the unit interval into a small set of probability intervals. An input sequence of discrete source symbols is mapped to a sequence of binary symbols and each of the binary symbols is assigned to one of the probability intervals. The binary symbols that are assigned to a particular probability interval are coded at a fixed probability using a simple code that maps a variable number of binary symbols to variable length codewords. The probability modeling is decoupled from the actual binary entropy coding. The coding efficiency of the probability interval partitioning entropy (PIPE) coding is comparable to that of arithmetic coding.
Key concepts: Huffman coding, Variable-length code, Shannon–Fano coding, Tunstall coding, Arithmetic coding, Context-adaptive binary arithmetic coding, Entropy encoding, Mathematics