Search
Patexia Research
Patent No. US 11082718
Issue Date Aug 3, 2021
Claim this patent
PDF Unavailable

Patent 11082718 - Method and device for coding image and method and device for decoding image > Claims

  • 1. A method for image coding, comprising: constructing a representative pixel set of a coding block within the coding block, and determining information for constructing the representative pixel set;constructing predicted values of pixel samples in the coding block, a generation manner for the predicted values comprising: constructing the predicted values of the pixel samples according to reconstructed pixel samples in an image region outside the coding block and pixel values of the representative pixel set in the coding block; andcoding and writing the information for constructing the representative pixel set and information for constructing the predicted values into a video bitstream;wherein, for the pixel samples in the coding block, the generation manner for the predicted values comprises:for the reconstructed pixel samples outside the coding block, calculating distances between the reconstructed pixel samples and part or all of representative pixel samples of the representative pixel set, and setting the representative pixel samples of which the distances meet a first predefined condition as optimal representative pixel samples; andunder the condition that distances between the optimal representative pixel samples and the reconstructed pixel samples meet a second predefined condition, setting numerical values of the optimal representative pixel samples as the predicted values, and when the distances between the optimal representative pixel samples and the reconstructed pixel samples do not meet the second predefined condition, setting the numerical values of the reconstructed pixel samples as the predicted values.
    • 2. The method as claimed in claim 1, wherein the coding block is a coding region of an image, and comprises at least one of: a Largest Coding Unit (LCU), a Coding Tree Unit (CTU), a Coding Unit (CU), a sub-region of the CU, a Prediction Unit (PU), a Transform Unit (TU), a pixel string, a pixel group, a pixel micro-block and a pixel strip; andthe image region outside the coding block comprises at least one of: an image region outside a variant of the coding block and an image region formed by pixels which have been partially coded.
    • 3. The method as claimed in claim 1, wherein the generation manner for the predicted values comprises one of: setting reconstructed pixel samples outside the coding block or variants of the reconstructed pixel samples as the predicted values; andsetting the pixel values of the representative pixel set or variants of the pixel values of the representative pixel set as the predicted values.
    • 4. The method as claimed in claim 1, wherein the predicted values of the pixel samples in the coding block are constructed by using at least one of the following coding parameters, comprising: positions of the reconstructed pixel samples in an image;the information for constructing the representative pixel set;a number of pixels of the representative pixel set;positions of current coding pixel samples in the coding block;types of pixel components of the coding block, wherein one pixel component or two pixel components of the pixel components select the pixel values of the representative pixel set as predicted values, and the other pixel components, except the one pixel component or the two pixel components, of the pixel components select pixel values of the reconstructed pixel samples as predicted values; andflags for the predicted values.
      • 5. The method as claimed in claim 4, wherein positions of the reconstructed pixel samples outside the coding block in an image are divided into at least one of rows and/or columns adjacent to the coding block; when the reconstructed pixel samples are positioned inside at least one of the rows and/or columns adjacent to the coding block, numerical values of the representative pixel samples of the representative pixel set are set as the predicted values; and when the reconstructed pixel samples are positioned outside at least one of the rows and/or columns adjacent to the coding block, numerical values of the reconstructed pixel samples are set as the predicted values;or, flags for the predicted values are directly obtained from the video bitstream or obtained by deduction according to at least one existing parameter in the video bitstream; the flags for the predicted values are used for indicating one of numerical values of the representative pixel samples of the representative pixel set and numerical values of the reconstructed pixel samples outside the coding block to be selected as the current coding pixel samples in a coding process; and positions of the flags in the video bitstream comprise one of: a VPS, an SPS, a PPS, a slice header, a CTU header, a CU header, a coding block header and positions before one or a group or string of current coding pixel samples in the coding block;or, determining the predicted values according to the number of pixels in the representative pixel set comprises: when the number of the pixels is smaller than a second predetermined threshold value, setting numerical values of the representative pixel samples of the representative pixel set as the predicted values; and when the number of the pixels is not smaller than the second predetermined threshold value, setting the numerical values of the reconstructed pixel samples as the predicted values.
    • 6. The method as claimed in claim 1, wherein the distances are errors between pixels, the first predefined condition is that the errors are minimum, and the second predefined condition is that the errors are smaller than a first predetermined threshold value.
    • 7. The method as claimed in claim 1, wherein setting the optimal representative pixel samples or the reconstructed pixel samples as the predicted values further comprises at least one of: writing the flags indicating the predicted values into the video bitstream;writing at least one condition parameter for constructing the predicted values into the video bitstream, wherein the pixel samples meeting the at least one condition parameter are determined as the predicted values.
      • 8. The method as claimed in claim 7, wherein writing the flags indicating the predicted values into the video bitstream comprises at least one of: writing flags used for indicating predicted values selected for a whole sequence into at least one of the following parameter sets: a Video Parameter Set (VPS), a Sequence Parameter Set (SPS) and a Picture Parameter Set (PPS); writing flags used for indicating predicted values selected for an image into at least one of a PPS and slice header; and writing flags used for indicating predicted values selected for a block layer into at least one of an LCU and a CU; or, writing the at least one condition parameter for construction of the predicted values into the video bitstream comprises at least one of: writing the at least one condition parameter selected for a whole sequence into at least one of the following parameter sets: a VPS, an SPS and a PPS; writing the at least one condition parameter selected for an image into at least one of the PPS and slice header; and writing the at least one condition parameter selected for a block layer into at least one of an LCU layer and a CU layer.
  • 9. A method for image decoding, comprising: parsing a bitstream to acquire at least one decoding parameter of a decoding block, wherein the at least one decoding parameter comprises at least one of the following parameters: at least one parameter for constructing a representative pixel set of the decoding block, and at least one parameter for constructing predicted values of pixel samples in the decoding block; andconstructing the representative pixel set, and determining predicted values, wherein a generation manner for the predicted values comprises: constructing the predicted values according to reconstructed pixel samples and pixel values of the representative pixel set,wherein, for the pixel samples in the decoding block, the generation manner for the predicted values comprises:for the reconstructed pixel samples outside the decoding block, calculating distances between the reconstructed pixel samples and part or all of representative pixel samples of the representative pixel set, and setting the representative pixel samples of which the distances meet a first predefined condition as optimal representative pixel samples; andunder the condition that distances between the optimal representative pixel samples and the reconstructed pixel samples meet a second predefined condition, setting numerical values of the optimal representative pixel samples as the predicted values, and when the distances between the optimal representative pixel samples and the reconstructed pixel samples do not meet the second predefined condition, setting the numerical values of the reconstructed pixel samples as the predicted values.
    • 10. The method as claimed in claim 9, wherein the decoding block is a decoding region in an image, and comprises at least one of: a Largest Coding Unit (LCU), a Coding Tree Unit (CTU), a Coding Unit (CU), a sub-region of the CU, a Prediction Unit (PU), a Transform Unit (TU), a pixel string, a pixel group, a pixel micro-block and a pixel strip; andan image region outside the decoding block comprises at least one of: an image region outside a variant of the decoding block and an image region formed by pixels which have been partially decoded.
    • 11. The method as claimed in claim 9, wherein the generation manner for the predicted values comprises one of: setting reconstructed pixel samples outside the decoding block or variants of the reconstructed pixel samples as the predicted values; andsetting the pixel values of the representative pixel set or variants of the pixel values of the representative pixel set as the predicted values.
    • 12. The method as claimed in claim 9, wherein the predicted values of the pixel samples in the decoding block are constructed by using at least one of the following coding parameters, comprising: positions of the reconstructed pixel samples in an image;the parameters for constructing the representative pixel set;a number of pixels of the representative pixel set;positions of current decoding pixel samples in the decoding block;types of pixel components of the decoding block, wherein one pixel component or two pixel components of the pixel components select the pixel values of the representative pixel set as predicted values, and the other pixel components, except the one pixel component or the two pixel components, of the pixel components select pixel numerical values of the reconstructed pixel samples as predicted values; andflags for the predicted values.
      • 13. The method as claimed in claim 12, wherein dividing the positions of the reconstructed pixel samples outside the decoding block in the image into at least one of rows and/or columns adjacent to the decoding block; when the reconstructed pixel samples are positioned in at least one of rows and/or columns adjacent to the decoding block, setting numerical values of the representative pixel samples of the representative pixel set as the predicted values; and when the reconstructed pixel samples are positioned outside at least one of rows and/or columns adjacent to the coding block, setting numerical values of the reconstructed pixel samples as the predicted values; or, parsing at least one of the following data units in the bitstream to obtain the flags for the predicted values, the data units comprising at least one of: a VPS, an SPS, a PPS, a slice header, a CTU header, a CU header, a coding block header and data units before one or a group or string of current coding pixel samples in a coding block, wherein the flags for the predicted values are directly obtained from the bitstream or obtained by deduction according to at least one existing parameter in the video bitstream, and the flags for the predicted values are used for indicating whether numerical values of the representative pixel samples of the representative pixel set or numerical values of the reconstructed pixel samples outside the decoding block are set as the predicted values;or, determining the predicted values according to the number of pixels in the representative pixel set comprises: when the number of the pixels is smaller than a second predetermined threshold value, setting numerical values of the representative pixel samples of the representative pixel set as the predicted values; and when the number of the pixels is not smaller than the second predetermined threshold value, setting the numerical values of the reconstructed pixel samples as the predicted values.
    • 14. The method as claimed in claim 9, wherein the distances are errors between pixels, the first predefined condition is that the errors are minimum, and the second predefined condition is that the errors are smaller than a first predetermined threshold value.
    • 15. The method as claimed in claim 9, wherein setting the optimal representative pixel samples or the reconstructed pixel samples as the predicted values comprises at least one of: parsing the flags in the bitstream, and setting values indicated by the flags as the predicted values; and,parsing at least one condition parameter in the bitstream, and setting pixel samples meeting the at least one condition parameter as the predicted values.
      • 16. The method as claimed in claim 15, wherein parsing the flags in the bitstream comprises at least one of: parsing a parameter set, comprising at least one of a Video Parameter Set (VPS), a Sequence Parameter Set (SPS) and a Picture Parameter Set (PPS), where flags used for indicating predicted values selected for a whole sequence are located; parsing at least one of a PPS and slice header where flags used for indicating predicted values selected for an image are located; and parsing at least one of an LCU layer and a CU layer where flags used for indicating predicted values selected for a block layer are located; or, parsing the at least one condition parameter in the bitstream comprises at least one of: parsing a parameter set where the at least one condition parameter selected for a sequence is located, wherein the parameter set comprises at least one of: a VPS, an SPS and a PPS; parsing at least one of the PPS and slice header where the at least one condition parameter selected for an image is located; and parsing at least one of an LCU layer and a CU layer where the at least one condition parameter selected for a block layer is located.
  • 17. A device for image coding, comprising: a first determination component, arranged to construct a representative pixel set of a coding block within the coding block and determine information for constructing the representative pixel set;a second determination component, arranged to construct predicted values of pixel samples of the coding block, a generation manner for the predicted values comprising: constructing the predicted values of the pixel samples according to reconstructed pixel samples in an image region outside the coding block and pixel values of the representative pixel set in the block; anda coding component, arranged to code and write the information for constructing the representative pixel set and information for constructing the predicted values into a bitstream;wherein, for the pixel samples in the coding block, the generation manner for the predicted values comprises:for the reconstructed pixel samples outside the coding block, calculating distances between the reconstructed pixel samples and part or all of representative pixel samples of the representative pixel set, and setting the representative pixel samples of which the distances meet a first predefined condition as optimal representative pixel samples; andunder the condition that distances between the optimal representative pixel samples and the reconstructed pixel samples meet a second predefined condition, setting numerical values of the optimal representative pixel samples as the predicted values, and when the distances between the optimal representative pixel samples and the reconstructed pixel samples do not meet the second predefined condition, setting the numerical values of the reconstructed pixel samples as the predicted values.
  • 18. A device for image decoding, comprising: a parsing component, arranged to parse a bitstream to acquire at least one decoding parameter of a decoding block, wherein the at least one decoding parameter comprises at least one of the following parameters: at least one parameter for constructing a representative pixel set of the decoding block, and at least one parameter for constructing predicted values of pixel samples in the decoding block; anda third determination component, arranged to construct the representative pixel set, and determine predicted values, wherein a generation manner for the predicted values comprises: constructing the predicted values according to reconstructed pixel samples and pixel values of the representative pixel set;wherein, for the pixel samples in the decoding block, the generation manner for the predicted values comprises:for the reconstructed pixel samples outside the decoding block, calculating distances between the reconstructed pixel samples and part or all of representative pixel samples of the representative pixel set, and setting the representative pixel samples of which the distances meet a first predefined condition as optimal representative pixel samples; andunder the condition that distances between the optimal representative pixel samples and the reconstructed pixel samples meet a second predefined condition, setting numerical values of the optimal representative pixel samples as the predicted values, and when the distances between the optimal representative pixel samples and the reconstructed pixel samples do not meet the second predefined condition, setting the numerical values of the reconstructed pixel samples as the predicted values.
Menu