Region-based information compaction as for digital images

Abstract

A method and apparatus for preserving the dynamic range of a relatively high dynamic range information stream, illustratively a high resolution video signal, subjected to a relatively low dynamic range encoding and/or transport process(es). A relatively high dynamic range information stream is subjected to a segmentation and remapping process whereby each segment is remapped to the relatively low dynamic range appropriate to the encoding and/or transport process(es) utilized. An auxiliary information stream includes segment and associated remapping information such that the initial, relatively high dynamic range information stream may be recovered in a post-encoding (i.e. decoding) or post-transport (i.e., receiving) process.

1. A method for encoding an information frame, comprising the steps of:
 dividing said information frame into a plurality of information regions, at least one of said information regions comprising at least one information parameter having associated with it a plurality of intra-region values bounded by upper and lower value limits defining a dynamic range of said information parameter; 
 determining, for each of said at least one of said information regions, a respective maximal value and a minimal value of said at least one information parameter; 
 remapping, for each of said at least one of said information regions and according to a single manipulation of the respective determined maximal and minimal values, at least one respective of said plurality of intra-region values of said at least one information parameter; and 
 encoding each of said information regions; 
 wherein said information frame comprises an image frame and said at least one information parameter comprises at least one of a luminance parameter and a chrominance parameter; 
 wherein said steps of encoding and determining produce, respectively, an encoded image stream and an associated dynamic range enhancement stream; and 
 wherein said step of remapping is performed in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       2. A method for encoding an information frame, comprising the steps of:
 dividing said information frame into a plurality of information regions, at least one of said information regions comprising at least one information parameter having associated with it a plurality of intra-region values bounded by upper and lower value limits defining a dynamic range of said information parameter; 
 determining, for each of said at least one of said information regions, a respective maximal value and a minimal value of said at least one information parameter; 
 remapping, for each of said at least one of said information regions and according to a single manipulation of the respective determined maximal and minimal values, at least one respective of said plurality of intra-region values of said at least one information parameter; and 
 encoding each of said information regions; wherein said plurality of information frames comprise image frames, said at least one information parameter comprises at least one of a luminance parameter and a chrominance parameter; 
 wherein said steps of encoding and determining produce, respectively, an encoded image stream and an associated dynamic range enhancement stream; 
 wherein said steps of dividing, determining, remapping and encoding are repeated for each of a plurality of information frames; and 
 wherein said step of remapping is performed in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       3. A method for decoding an encoded information frame represented by a plurality of encoded information regions within an encoded information stream, where at least one of said plurality of encoded information regions comprises at least one information parameter having associated with it a plurality of remapped intra-region values, said method comprising the steps of:
 decoding each of said plurality of encoded information regions to form a corresponding plurality of decoded information regions, said decoded information regions representing a decoded information frame; 
 extracting, from a dynamic range enhancement stream associated with said encoded information stream, respective maximal and minimal values for each of said at least one information parameter having associated with it a plurality of remapped intra-region values; 
 inverse remapping, according to a single manipulation of said respective maximal and minimal values, each of said at least one information parameter of said at least one information regions having associated with it a respective plurality of remapped intra-region values; and 
 demultiplexing a transport stream to recover said encoded information stream and said dynamic range enhancement stream; wherein said encoded information stream and said dynamic range enhancement stream associated with said encoded information stream comprises respective portions of said transport stream; 
 wherein said step of demultiplexing comprises retrieving, from a private data section of said transport stream, said dynamic range enhancement stream; 
 wherein said decoded information frame comprises an image frame and said at least one information parameter comprises at least one of a luminance parameter and a chrominance parameter; and 
 wherein said step of inverse remapping is performed in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       4. A method for decoding an encoded information frame represented by a plurality of encoded information regions within an encoded information stream, where at least one of said plurality of encoded information regions comprises at least one information parameter having associated with it a plurality of remapped intra-region values, said method comprising the steps of:
 decoding each of said plurality of encoded information regions to form a corresponding plurality of decoded information regions, said decoded information regions representing a decoded information frame; 
 extracting, from a dynamic range enhancement stream associated with said encoded information stream, respective maximal and minimal values for each of said at least one information parameter having associated with it a plurality of remapped intra-region values; and 
 inverse remapping, according to a single manipulation of said respective maximal and minimal values, each of said at least one information parameter of said at least one information regions having associated with it a respective plurality of remapped intra-region values; 
 wherein said steps of decoding, extracting and inverse remapping are repeated for each of a plurality of information frames within said encoded information stream and said dynamic range enhancement stream associated with said encoded information stream; 
 wherein said decoding step comprises compression decoding said encoded information frame of said information stream; and 
 wherein said step of inverse remapping is performed in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       5. An apparatus for decoding an encoded information frame represented by a plurality of encoded information regions within an encoded information stream, where at least one of said plurality of encoded information regions comprises at least one information parameter having associated with it a plurality of remapped intra-region values, said apparatus comprising:
 a decoder, for decoding each of said plurality of encoded information regions to form a corresponding plurality of decoded information regions, said decoded information regions representing a decoded information frame; and 
 an inverse map and scale unit, for extracting, from a dynamic range enhancement stream associated with said encoded information stream, respective maximal and minimal values for each of said at least one information parameter having associated with it a plurality of remapped intra-region values, and for inverse remapping, according to a single manipulation of said respective maximal and minimal values, each of said at least one information parameter of said at least one of said plurality of encoded information regions having associated with it said plurality of remapped intra-region values; 
 wherein said decoded information frame comprises an image frame and said at least one information parameter comprises at least one of a luminance parameter and a chrominance parameter; and 
 wherein said inverse map and scale unit extracts said respective maximal and minimal values in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       6. An apparatus for decoding an encoded information frame represented by a plurality of encoded information regions within an encoded information stream, where at least one of said plurality of encoded information regions comprises at least one information parameter having associated with it a plurality of remapped intra-region values, said apparatus comprising:
 a decoder, for decoding each of said plurality of encoded information regions to form a corresponding plurality of decoded information regions, said decoded information regions representing a decoded information frame; and 
 an inverse map and scale unit, for extracting, from a dynamic range enhancement stream associated with said encoded information stream, respective maximal and minimal values for each of said at least one information parameter having associated with it a plurality of remapped intra-region values, and for inverse remapping, according to a single manipulation of said respective maximal and minimal values, each of said at least one information parameter of said at least one of said plurality of encoded information regions having associated with it said plurality of remapped intra-region values; 
 wherein said encoded information stream and said dynamic range enhancement stream associated with said encoded information stream represent a plurality of encoded information frames; 
 wherein said apparatus processes each of said plurality of encoded information frames to produce a corresponding plurality of decoded information frames; 
 wherein said plurality of decoded information frames comprise image frames and said at least one information parameter comprises at least one of a luminance parameter and a chrominance parameter; and 
 wherein said inverse map and scale unit extracts said respective maximal and minimal values in accordance with the following linear equation:
   TP=[OP*(TR/OR)+0.5] 
 
 where: TP=Target Pixel; OP=Original Pixel; TR=Target Range; and OR=Original Range. 
 
     
     
       7. A method for remapping a value of a parameter of an information element from a first representation system to a second representation system, comprising:
 dividing a collection of information elements into regions; 
 determining a region of the regions, wherein the information element is within the region; 
 determining, among the information elements within the region, a maximum value of the parameter; 
 determining, among the information elements within the region, a minimum value of the parameter; 
 subtracting the minimum value from the value of the parameter of the information element to yield a first difference; 
 subtracting the minimum value from the maximum value to yield a second difference; 
 dividing a dynamic range of values of the parameter associated with the second representation system by the second difference to yield a quotient; 
 multiplying the quotient by the first difference to yield a product; 
 adding the product to one half to yield a sum; and 
 rounding the sum to a nearest integer that is less than or equal to the sum to yield a remapped value of the parameter of the information element. 
 
     
     
       8. The method of  claim 7 , wherein the second difference is raised to a power before the dividing the dynamic range of values of the parameter associated with the second representation system by the second difference, the first difference is raised to the power before the multiplying the quotient by the first difference, a function is determined before the adding the product to one half to yield the sum, and the product is a variable of the function. 
     
     
       9. The method of  claim 7 , wherein the information element is a pixel. 
     
     
       10. The method of  claim 9 , wherein the collection of information elements is an image frame. 
     
     
       11. The method of  claim 7 , further comprising performing, for each remaining information element within the region, the subtracting the minimum value from the value of the parameter of the information element, the subtracting the minimum value from the maximum value, the dividing the dynamic range of values of the parameter associated with the second representation system by the second difference, the multiplying the quotient by the first difference, the adding, and the rounding. 
     
     
       12. The method of  claim 11 , further comprising performing, for each remaining region of the regions, the determining, among the information elements within the region, the maximum value of the parameter and the determining, among the information elements within the region, the minimum value of the parameter to yield a collection of maximum values and a collection of minimum values. 
     
     
       13. The method of  claim 12 , further comprising performing, for each information element within the each remaining region, the subtracting the minimum value from the value of the parameter of the information element, the subtracting the minimum value from the maximum value, the dividing the dynamic range of values of the parameter associated with the second representation system by the second difference, the multiplying the quotient by the first difference, the adding, and the rounding. 
     
     
       14. The method of  claim 13 , further comprising encoding the regions to yield encoded regions. 
     
     
       15. The method of  claim 14 , further comprising transporting the encoded regions, the collection of maximum values, and the collection of minimum values. 
     
     
       16. The method of  claim 13 , further comprising decoding the regions. 
     
     
       17. A method for remapping a value of a parameter of an information element from a first representation system to a second representation system, comprising:
 dividing a dynamic range of values of the parameter associated with the second representation system by a dynamic range of values of the parameter associated with the first representation system to yield a quotient; 
 multiplying the quotient by the value of the parameter of the information element to yield a product; 
 adding the product to one half to yield a sum; and 
 rounding the sum to a nearest integer that is less than or equal to the sum to yield a remapped value of the parameter of the information element. 
 
     
     
       18. An apparatus for remapping a value of a parameter of an information element from a first representation system to a second representation system, comprising:
 a receiver configured to receive a collection of information elements; and 
 a processor configured to divide the collection of information elements into regions, to determine a region of the regions, wherein the information element is within the region, to determine, among the information elements within the region, a maximum value of the parameter, to determine, among the information elements within the region, a minimum value of the parameter, to subtract the minimum value from the value of the parameter of the information element to yield a first difference, to subtract the minimum value from the maximum value to yield a second difference, to divide a dynamic range of values of the parameter associated with the second representation system by the second difference to yield a quotient, to multiply the quotient by the first difference to yield a product, to add the product to one half to yield a sum, and to round the sum to a nearest integer that is less than or equal to the sum to yield a remapped value of the parameter of the information element. 
 
     
     
       19. The apparatus of  claim 18 , wherein the processor is further configured to raise the second difference to a power before dividing the dynamic range of values of the parameter associated with the second representation system by the second difference, to raise the first difference to the power before multiplying the quotient by the first difference, and to determine a function before adding the product to one half to yield the sum, wherein the product is a variable of the function. 
     
     
       20. The apparatus of  claim 18 , further comprising an encoder configured to encode the regions to yield encoded regions. 
     
     
       21. The apparatus of  claim 20 , further comprising a transporter configured to transport the encoded regions, a collection of maximum values, and a collection of minimum values. 
     
     
       22. The apparatus of  claim 18 , further comprising a decoder configured to decode the regions. 
     
     
       23. An apparatus for remapping a value of a parameter of an information element from a first representation system to a second representation system, comprising a processor configured to divide a dynamic range of values of the parameter associated with the second representation system by a dynamic range of values of the parameter associated with the first representation system to yield a quotient, to multiply the quotient by the value of the parameter of the information element to yield a product, to add the product to one half to yield a sum, and to round the sum to a nearest integer that is less than or equal to the sum to yield a remapped value of the parameter of the information element. 
     
     
       24. In a system for encoding an information stream, a method comprising:
 dividing the information stream into a plurality of information regions;   identifying, for at least one of the plurality of information regions, a maximal value and a minimal value of at least one information element associated with the least one of the plurality of information regions;   mapping the at least one information element according to the identified maximal and minimal values associated with the respective information region, wherein said mapping comprises determining a target information element as a function of an original information element, a target range, and said maximal and minimal values, wherein the difference between the maximal and minimal values is used to map the original information element to the target range to generate a mapped information element;   encoding the mapped information element to produce an encoded information stream; and   associating said identified maximal and minimal values with the at least one of the plurality of information regions to produce a map identification stream, wherein said map identification stream includes information sufficient to substantially recover said identified maximal and minimal values associated with said mapped at least one information element.   
     
     
       25. The method of claim 24, wherein said mapping further comprises using a combination of linear and non-linear functions. 
     
     
       26. The method of claim 24, wherein said mapping further comprises using a polynomial segment. 
     
     
       27. The method of claim 24, wherein said mapping further comprises using a tabulated function comprising an indexable array of values. 
     
     
       28. A method for encoding an information stream, said method comprising:
 receiving a plurality of information frames, each of said plurality of information frames comprising a plurality of information elements, each of said plurality of information elements associated with at least one information element parameter, each of the at least one information element parameters having a value between a lower limit and an upper limit, said upper and lower limits defining a dynamic range of said at least one information element parameter;   defining, for each of the plurality of information frames, a plurality of information regions, each of the plurality of information regions being associated with one or more respective information elements;   identifying, for each of the plurality of information regions, a maximal value and a minimal value of at least one information element parameter associated with said plurality of information elements of said region;   remapping, for each of said plurality of information regions, the at least one information element parameter of each of said plurality of information elements of said region according to the identified maximal and minimal value parameters associated with the respective region, wherein the difference between the maximal and minimal value parameters is used to remap the at least one of said plurality of information element parameters to a target range;   encoding each remapped information region of said plurality of information regions to produce an encoded information stream; and   associating said identified maximal and minimal values with each remapped information region of said plurality of information regions to produce a map identification stream, wherein said map identification stream includes information sufficient to substantially recover said identified maximal and minimal value parameters associated with said remapped at least one information element parameter for each of said plurality of information regions.   
     
     
       29. A method for encoding an information stream comprising a plurality of information frames, said method comprising:
 receiving the plurality of information frames;   for at least one of the plurality of information frames, dividing the at least one of the plurality of information frames into regions according to at least one criterion for rescaling a range of one or more parameters of interest for at least one of the regions as compared to a range of the one or more parameters of interest of the at least one of the plurality of information frames;   identifying a target range for said one or more parameters of interest, the target range being different from the range of the one or more parameters of interest; and   remapping, for at least one of said regions, the one or more parameters of interest to the identified target range, the remapped parameters being bounded by the target range.   
     
     
       30. The method of claim 29, further comprising encoding the remapped regions to produced a compressed information stream. 
     
     
       31. The method of claim 30, further comprising multiplexing the encoded remapped regions with information for recovering the regions.