Entropy Coding Using Pre-Defined, Fixed CDFS
AssigneeGoogle Inc.
InventorsJoseph Young, In Suk Chong, Debargha Mukherjee
Life of this patent
Estimating the term…
Abstract
A predictor value corresponding to a value of a sequence of values is determined. The predictor value is generated by an adaptable predictor function based on a statistical distribution of the sequence of values. A probability model is selected from a set of pre-defined probability models based on the predictor value. A symbol associated with the value is entropy decoded from a compressed bitstream using the probability model.
Claims20
A method, comprising:
- determining a predictor value corresponding to a value of a sequence of values, the predictor value being generated by an adaptable predictor function based on a statistical distribution of the sequence of values;
- selecting, based on the predictor value, a probability model from a set of pre-defined probability models;
- and
- entropy decoding, from a compressed bitstream and using the probability model, a symbol associated with the value.
A device, comprising:
- a processor, the processor configured to execute instructions to:
- determine a predictor value corresponding to a value of a sequence of values;
- select, based on the predictor value, a probability model from a set of pre-defined probability models;
- and
- entropy decode, from a compressed bitstream and using the probability model, a symbol associated with the value.
A non-transitory computer-readable storage medium storing a compressed bitstream, wherein, the compressed bitstream, processed by a processor, causes the processor to perform operations comprising:
- determining a predictor value corresponding to a value of a sequence of values;
- selecting, based on the predictor value, a probability model from a set of pre-defined probability models;
- and
- entropy decoding, from the bitstream and using the probability model, a symbol associated with the value.
Dependencies are inferred from the claim text.