摘要:
A method for encoding an input sequence of symbols. The method includes, sequentially, for each symbol in the input sequence, determining an estimated probability for that symbol based on a context model, identifying a codeword associated with a sequence of symbols resulting from appending that symbol to a previous sequence of symbols associated with that estimated probability, using an encoding tree associated with that estimated probability, and storing the codeword in a buffer element of a first-in-first-out buffer, wherein the buffer element is associated with the estimated probability. Stored codewords are output from the buffer in a first-in-first-out order.
摘要:
A method for encoding an input sequence of symbols as a plurality of codewords including primary codewords and secondary codewords, and a method of decoding the plurality of codewords. The codewords are output as a bitstream including an associated codeword type indicator, which specifies whether its associated plurality of consecutive codewords includes at least one secondary codeword and, if so, the location of the first of the at least one secondary codewords. The decoder selects between a primary codeword decoding tree and a secondary codeword decoding tree, and decodes one of the plurality of codewords using the selected decoding tree.
摘要:
Methods of encoding and decoding for video data are describe in which significance maps are encoded and decoded using non-spatially-uniform partitioning of the map into parts, wherein the bit positions within each part are associated with a given context. Example partition sets and processes for selecting from amongst predetermined partition sets and communicating the selection to the decoder are described.
摘要:
Methods and devices for encoding and decoding that involve sorting bins according to their respective estimated probabilities to form subsequences, each subsequence having an associated estimated probability. Subsequences are encoded to form codewords. Ordered sets of phrases of known length are then formed from the codewords. Each first of the phrases in a set contains at least part of one codeword. The first phrase has an associated estimated probability and the probability estimates associated with each of the other phrases in the set are determined based upon the probability estimate associated with the first phrase, which permits the phrases to be decoded in parallel.
摘要:
A method for parallel context modeling through reordering the bits of an input sequence to form groups of bits in accordance with a context model-specific reordering schedule. The reordering schedule is developed such that the groups of bits are formed to satisfy two conditions: first, that the context for each of the bits in a group of bits is different from the context of each of the other bits in that group, and the context of each of the bits in that group is determined independently from each of the other bits in that group. The parallel context modeling may be used in encoding or decoding operations.
摘要:
A method for encoding an input sequence of symbols. The method includes, sequentially, for each symbol in the input sequence, determining an estimated probability for that symbol based on a context model, identifying a codeword associated with a sequence of symbols resulting from appending that symbol to a previous sequence of symbols associated with that estimated probability, using an encoding tree associated with that estimated probability, and storing the codeword in a buffer element of a first-in-first-out buffer, wherein the buffer element is associated with the estimated probability. Stored codewords are output from the buffer in a first-in-first-out order.
摘要:
Methods and systems are described for combining sources in a data compression system. In a system in which a context model results in the production of multiple sources or symbols, each source being associated with a probability estimate, sources may be combined by defining another, reduced size, set of probabilities. The new set of probabilities may be a subset of the predefined set of probabilities provided by the context model. Minimizing relative entropy may be a basis upon which to define a mapping of predefined probabilities to the new set of probabilities. An input sequence that was modeled using the context model may then be entropy encoded and entropy decoded using the new set of probabilities based on a mapping between the new probabilities and the predefined probabilities.
摘要:
In some implementations, a method for encoding data comprising a matrix of elements for scaling transform coefficients before quantization of the scaled transform coefficients includes generating a sequence of values from the elements of the matrix according to a predetermined order. A plurality of adjacent values in the sequence is generated from respective elements of the matrix. A representation of the data is encoded based at least in part on encoding repeated instances of a specified series of two or more values in the sequence as a corresponding symbol not appearing in the sequence.
摘要:
A method is presented for entropy coding data using an entropy coder to encode an input sequence. A context model is used to determine the context of each symbol and a probability estimation is made for each symbol. A method is presented for revising the coding order to be context-based, grouping symbols consecutively on the basis that they have a common context. A method is presented for entropy decoding a bitstream of encoded data encoded using a context-based coding order.
摘要:
A method for encoding an input sequence of symbols to produce a bitstream and a method of decoding the bitstream to generate a reconstructed binary sequence. Encoding employs an encoding tree having primary codewords associated with leaf nodes and secondary codewords associated with internal nodes. A flush event may cause output of secondary codewords. A context model is used to select an encoding tree corresponding to an estimated probability at the encoder. The same context model is used by the decoder to select a decoding tree. The decoder interleaves bits from decoded bit sequences associated with different estimated probabilities based on the context model.