1. An article of manufacture having a machine-readable storage medium including associated instructions, wherein the instructions, when executed, result in the machine performing:
subdividing at least one basic block of a target program into a plurality of regions;
instrumenting each region to transform a current region signature within a signature register into a successor region signature at an end of each of the plurality of regions;
instrumenting the basic block to include a signature checking region, the signature checking region to verify that contents of the signature register match a signature of the signature checking region; and
instrumenting the basic block to update the signature register with a successor basic block signature at an end of the basic block; and
inserting a transformation instruction to combine the contents of the signature register with a constant value to form a control flow signature during runtime of the target program.
2. The article of manufacture of claim 1, wherein the machine accessible medium further comprises data, which when accessed, results in the machine further performing:
subdividing at least one basic block of a target program into at least a first region and a successor second region;
instrumenting the basic block to transform a signature of the basic block, stored within a signature register, into a signature of the first region prior to the first region;
instrumenting the basic block to verify that contents of the signature register match the signature of the first region prior to the first region; and
instrumenting the basic block to update the signature register with a signature of the successor second region between the first region and the successor second region.
3. The article of manufacture of claim 2, wherein instrumenting the at least one basic block further results in the machine performing:
inserting a conditional instruction to update the signature register if the at least one basic block includes a conditional branch instruction.
4. The article of manufacture of claim 2, wherein instrumenting the at least one basic block further results in the machine performing:
instrumenting the at least one basic block to identify a dynamic target address if the at least one basic block includes a dynamic branch instruction;
instrumenting the basic block to map the dynamic target address to a target basic block signature; and
inserting a conditional instruction to transform contents of the signature register to identify a target successor basic block of the at least one basic block, according to the dynamic target address.
5. The article of manufacture of claim 2, wherein instrumenting the basic block to verify further results in the machine performing:
inserting a transformation instruction to combine the contents of the signature register with a constant value to form a region signature;
inserting an instruction within the basic block to compare the region signature to the signature of the first region; and
inserting an instruction within the basic block subsequent to a compare instruction, to invoke an error handler if the region signature does not match the signature of the first region.
6. A system comprising:
a processor;
a memory controller coupled to the processor; and
a memory coupled to the memory controller, the memory controller including a binary translator to instrument at least one basic block of a target program to include signature update code to update a signature register with a current basic block signature at a beginning of the basic block, to instrument the basic block to verify that contents of the signature register match a signature of the basic block subsequent to the signature update code and to instrument the basic block to update the signature register with a successor basic block signature at an end of the basic block;
wherein the translator is further to subdivide the at least one basic block of the target program into at least a first region and a successor second region, to instrument the basic block to transform a signature of the basic block, stored within a signature register, into a signature of the first region prior to the first region, and to instrument the basic block to verify that contents of the signature register match the signature of the first region prior to the first region.
7. The system of claim 6, wherein the translator is further to instrument the basic block to update the signature register with a signature of the successor second region between the first region and the successor second region.
8. The system of claim 6, wherein the translator is further to insert a transformation instruction to combine the contents of the signature register with a constant value to form a region signature, to insert an instruction within the basic block to compare the region signature to the first region signature, and to insert an instruction within the basic block subsequent to a compare instruction, to invoke an error handler if the region signature does not match the first region signature.
9. The system of claim 6, wherein the translator is further to instrument the successor basic block to update the signature register with a next successor basic block signature of a next successor block of the successor basic block at an end of the successor basic block.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
1-10. (canceled)
11. A method for correcting gradations on the basis of Retinex theory in a display device, comprising:
obtaining each assignment luminance value to be assigned to each processing unit gradation, on the basis of the maximum and minimum luminance values of each display device;
obtaining each discernible luminance difference corresponding to the each assignment luminance value on the basis of data on discernible luminance differences corresponding to luminance values;
obtaining thresholds of an epsilon filter such that the thresholds have a monotone increasing relationship with the assignment luminance values, on the basis of the obtained discernible luminance differences; and
when input image data of a predetermined region is provided to the epsilon filter, correcting gradations using the threshold of the epsilon filter corresponding to a center representative value of the input image data.
12. The method for correcting gradations of claim 11, wherein the each assignment luminance value is obtained on the basis of the maximum and minimum luminance values, as well as ambient light.
13. The method for correcting gradations of claim 12, wherein a formula representing a relationship between the processing unit gradations and the thresholds of the epsilon filter is dynamically changed based on ambient light detected by ambient light detection means.
14. A device for correcting gradations on the basis of Retinex theory in a display device, comprising:
an extraction unit configured to, when receiving a predetermined region of input image data to be inputted to an epsilon filter, extract a center representative value located around the center of an adjacent pixel region, the center representative value being to be used when the epsilon filter performs a calculation;
a threshold determination unit storing correspondences between the center representative value and the threshold of the epsilon filter and configured to, when receiving the center representative value, determine a corresponding threshold on the basis of the correspondences;
an epsilon filter configured to smooth the inputted predetermined region of input image data using the threshold determined by the threshold determination unit; and
a gradation correction unit configured to correct gradations using an output value of the epsilon filter as an illumination light component.
15. A device for determining a threshold of an epsilon filter in Retinex processing, comprising:
an extraction unit configured to extract a center representative value located around the center of an adjacent pixel region, the center representative value being to be used when the epsilon filter performs a calculation; and
a threshold determination unit storing correspondences between the center representative value and the threshold of the epsilon filter and configured to, when receiving the center representative value, determine a corresponding threshold on the basis of the correspondences, wherein
the correspondences between the center representative value and the threshold of the epsilon filter represent a monotone increasing relationship determined based on a discernible luminance difference corresponding to a particular luminance value.
16. The device for determining a threshold of an epsilon filter of claim 15, wherein the center representative value is luminance of a center pixel in the adjacent pixel region.
17. The device for determining a threshold of an epsilon filter of claim 15, wherein the correspondences between the center representative value and the threshold of the epsilon filter are represented by a formula \u03b5=(center representative value*\u03b1)+\u03b2 where \u03b1 and \u03b2 each represent a decimal fraction between 0 and 1, and the center representative value is normalized to 0 to 1.
18. The device for determining a threshold of an epsilon filter of any one of claim 15, further comprising an ambient light detection unit configured to detect ambient light, wherein the threshold determination unit stores a plurality of correspondences between the center representative value and the threshold of the epsilon filter, the correspondences corresponding to values of the ambient light, and selects one of the correspondences on the basis of the detected ambient light and outputs a corresponding threshold.
19. A method for determining a threshold of an epsilon filter in Retinex processing, comprising:
obtaining each discernible luminance difference corresponding to each processing unit gradation with respect to each assignment luminance value assigned to the each processing unit gradation, on the basis of data on discernible luminance differences corresponding to luminance values; and
determining the assignment luminance values and the threshold in such a manner that the assignment luminance values and the threshold have a monotone increasing relationship with a center representative value provided to the epsilon filter, on the basis of the obtained discernible luminance differences.
20. A method for correcting gradations on the basis of Retinex theory in a display device, comprising:
obtaining each assignment luminance value to be assigned to each processing unit gradation, on the basis of the maximum and minimum luminance values of each display device;
obtaining each discernible luminance difference corresponding to the each obtained assignment luminance value on the basis of data on discernible luminance differences corresponding to luminance values;
obtaining each normalized discernible luminance difference corresponding to the each processing unit gradation by normalizing the obtained discernible luminance differences using a difference between the maximum and minimum values of the display device;
obtaining thresholds of an epsilon filter corresponding to the respective processing unit gradations on the basis of the normalized discernible luminance differences; and
when input image data of a predetermined region is provided to the epsilon filter, correcting gradations using the threshold of the epsilon filter corresponding to a center representative value of the input image data.