Lossless compression of fragmented image data
Lossless compression of fragmented image data is disclosed. In some embodiments, a stream of information comprising data elements having statistical characteristics is received. An encoded output is produced by an encoder comprising a data compressor that implements a variable length code that is adapted to the statistical characteristics of the data elements. The output and information from which the variable length code can be derived are stored.
1. A system, comprising:
a processor configured to:
obtain a data input sequence comprising data elements; and generate a compressed output corresponding to the data elements,
wherein:
the data input sequence corresponds to information that is received from one or more channels,
the compressed output is generated based at least in part on (i) performing one or more transformations with respect to the data input sequence to exploit one or more continuity properties of information received from a plurality of the one or more channels or based at least in part on one or more continuity properties of information received within one of the one or more channels, (ii) use of at least one of a lossless coding, a Huffman code, or a variable length code, and (iii) a grouping of the data elements into intervals, the data elements being grouped based at least in part on one or more properties of the data elements and according to one or more thresholds, the one or more thresholds being determined according to one or more of a distribution of values in corresponding intervals and a frequency of values in the corresponding intervals; and
a memory that stores the compressed output.
2. The system of claim 1 , wherein the processor is further configured to:
convert each of the data elements into a level and value pair representation according to the determined one or more thresholds, wherein the level specifies high order bits and the value specifies low order bits; and
encode the levels of the data elements using a variable length code adapted to the one or more properties of levels of the data elements.
3. The system of claim 2 , wherein the determined one or more thresholds are data type adaptive and depend on the kind of data comprising the data input sequence.
4. The system of claim 2 , wherein the determined one or more thresholds are data instance adaptive and depend on the actual data comprising the data input sequence.
5. The system of claim 2 , wherein the compressed output comprises information for reconstructing the variable length code, the encoded levels, and unencoded value bits.
6. The system of claim 2 , wherein the level and value pair representation correspond to a signed representation.
7. The system of claim 2 , wherein the level and value pair representation correspond to an unsigned representation.
8. The system of claim 2 , wherein the level and value pair representation correspond to an unsigned representation and signs of the data elements are represented separately in the compressed output.
9. The system of claim 2 , wherein the variable length code adapted to one or more properties of the levels of the data elements is adapted to the frequencies of the levels of the data elements.
10. The system of claim 2 , wherein in a level and value pair representation of a data element, the level component specifies an interval in which the data element falls and the value component specifies a particular integer in that interval.
11. The system of claim 2 , wherein the variable length code comprises a Huffman code.
12. The system of claim 2 , wherein to encode the levels comprises to encode each level using information about a previous level in the sequence.
13. The system of claim 2 , wherein to encode the levels comprises to divide the levels into groups and generate a separate variable length code for each group.
14. The system of claim 1 , wherein the one or more transforms increases compressibility of the information received from the one of the one or more channels.
15. The system of claim 1 , wherein the compressed output comprises a lossless compression of the data input sequence.
16. The system of claim 1 , wherein the data input sequence correspond to a fragment of an image.
17. The system of claim 16 , wherein the fragment is independently compressed and decompressed.
18. A method, comprising:
obtaining a data input sequence comprising data elements;
generating a compressed output corresponding to the data elements; and
storing the compressed output,
wherein:
the data input sequence corresponds to information that is received from one or more channels,
the compressed output is generated based at least in part on (i) performing one or more transformations with respect to the data input sequence to exploit one or more continuity properties of information received from a plurality of the one or more channels or based at least in part on one or more continuity properties of information received within one of the one or more channels, (ii) use of at least one of a lossless coding, a Huffman code, or a variable length code, and (iii) a grouping of the data elements into intervals, the data elements being grouped based at least in part on one or more properties of the data elements and according to one or more thresholds, the one or more thresholds being determined according to one or more of a distribution of values in corresponding intervals and a frequency of values in the corresponding intervals.
19. A computer program product embodied in a non-transitory computer readable storage medium and comprising computer instructions for:
obtaining a data input sequence comprising data elements;
generating a compressed output corresponding to the data elements; and
storing the compressed output,
wherein:
the data input sequence corresponds to information that is received from one or more channels,
the compressed output is generated based at least in part on (i) performing one or more transformations with respect to the data input sequence to exploit one or more continuity properties of information received from a plurality of the one or more channels or based at least in part on one or more continuity properties of information received within one of the one or more channels, (ii) use of at least one of a lossless coding, a Huffman code, or a variable length code, and (iii) a grouping of the data elements into intervals, the data elements being grouped based at least in part on one or more properties of the data elements and according to one or more thresholds, the one or more thresholds being determined according to one or more of a distribution of values in corresponding intervals and a frequency of values in the corresponding intervals.
20. The method of claim 18 , wherein the generating the compressed output corresponding to the data elements comprises:
transforming the data input sequence based at least in part on one or more continuity properties of information received across a plurality of the one or more channels or based at least in part on one or more continuity properties of information received within one of the one or more channels, wherein the transforming the data input sequence comprises performing one or more transforms on the data elements.
21. The method of claim 20 , wherein the transformed data input sequence comprises data elements having statistical characteristics.
22. The method of claim 21 , compressing the transformed data input sequence according to one or more of statistical characteristics of the data elements, wherein the compressed output is generated based at least in part on the compressing the transformed data input sequence.