Disclosed are an apparatus and a method of encoding/decoding a video, particularly a method and an apparatus for storing a quantization parameter differential value in a largest coding unit (LCU) based on quadtree splitting and adaptively predicting a quantization parameter value based on context information on neighboring CUs. Quadtree-based quantization parameter encoding and decoding methods and apparatuses effectively show information on a block having a quantization parameter differential value based on splitting information on a CU and adaptively predict a quantization parameter value using context information including a block size, block partition and a quantization parameter of a neighboring CU.
Legal claims defining the scope of protection, as filed with the USPTO.
determining a parameter recorded unit for which dequantization parameter information is signaled, wherein the parameter recorded unit is split in a quadtree structure; determining a parameter prediction unit used for deriving a predicted dequantization parameter, wherein the parameter prediction unit comprises the parameter recorded unit; deriving the predicted dequantization parameter for the parameter prediction unit, wherein the predicted dequantization parameter is derived using an average value of dequantization parameters of a block neighboring to a left of the parameter prediction unit and a block neighboring above the parameter prediction unit; deriving a differential dequantization parameter for the parameter recorded unit based on the dequantization parameter information; deriving the dequantization parameter using a sum of the predicted dequantization parameter and the differential dequantization parameter; and performing dequantization using the dequantization parameter. . A method for decoding a video by a decoding apparatus using derivation of a dequantization parameter, the method comprising:
claim 1 wherein the quantization parameter flag indicates whether information with which the dequantization parameter is predicted for coding unit is present or not. . The method of, wherein the parameter prediction unit is a block which has a size same with one of a coding unit size from size of smallest coding unit to size of largest coding unit when a quantization parameter flag has a value of true, and
claim 1 wherein the parameter recorded unit is split into blocks of lower depth in a quadtree structure when the split flag is equal to 1, and the parameter recorded unit is not split further when the split flag is equal to 0. . The method of, wherein splitting information is a split flag, and
claim 1 wherein the split flag equal to 0 indicates that the differential dequantization parameter is signaled for the parameter recorded unit and the split flag equal to 1 indicates that the parameter recorded unit is split into sub-blocks and the differential dequantization parameter is not signaled for the parameter recorded unit, and wherein split flags for the sub-blocks are signaled when the split flag is equal to 1. . The method of, wherein splitting information is a split flag,
determine a unit image block for parameter recording for which dequantization parameter information is signaled, wherein the unit image block for parameter recording is split in a quadtree structure; determine a unit image block for parameter predicting, used for deriving a predicted dequantization parameter, wherein the unit image block for parameter predicting comprises the unit image block for parameter recording; one or more modules to: derive the predicted dequantization parameter for the unit image block for parameter predicting, wherein the predicted dequantization parameter is derived using an average value of dequantization parameters of a block neighboring to e left of the unit image block and a block neighboring above the unit image block; derive a differential dequantization parameter for the unit image block for parameter recording based on the dequantization parameter information; derive the dequantization parameter using a sum of the predicted dequantization parameter and the differential dequantization parameter; and perform dequantization using the dequantization parameter. a dequantization parameter derivation module to: 5. An apparatus for decoding a video using derivation of a dequantization parameter, the apparatus comprising:
claim 5 wherein the quadtree structure splits a block into four sub-blocks with half horizontal size. 6. The apparatus of:
Complete technical specification and implementation details from the patent document.
More than one Reissue Application has been filed for the reissue of the U.S. Pat. No. 9,066,098, issued on Jun. 23, 2015. The reissue applications are application Ser. No. 14/815,325, filed on Jul. 31, 2015 (now Pat. No. RE47,465, issued on Jun. 25, 2019); application Ser. No. 14/815,532, filed on Jul. 31, 2015 (now Pat. No. RE46,678, issued on Jan. 16, 2018); application Ser. No. 16/407,669, filed on May 9, 2019 (Now Pat. No. RE49,330, issued on Dec. 6, 2022); and application number 17/970,479 (the present application), filed on Oct. 20, 2022.This applicationclaims the benefit as a Reissue Continuation Application of U.S. application Ser. No. 16/407,669, filed May 9, 2019, which claims the benefit as a Reissue Continuation Application of U.S. application Ser. No. 14/815,325, filed Jul. 31, 2015, which is a Reissue Application from U.S. Pat. No. 9,066,098 issued on Jun. 23, 2015 and filed Dec. 30, 2014, whichis a Continuation Application of U.S. application Ser. No. 14/096,211, filed on Dec. 4, 2013(now U.S. Pat. No. 8,964,834, issued on Feb. 24, 2015), which is a Continuation Application of PCT Application No. PCT/KR2012/004881, filed on Jun. 20, 2012 and published as WO 2012/177051 on Dec. 27, 2012, which claims priority to Korean Patent Application No. 10-2011-0060058, filed on Jun. 21, 2011 in the Korean Intellectual Property Office, the entire disclosures of which are incorporated herein by reference for all purposes.Applicants hereby rescind any disclaimer of claim scope in the parent application(s) or the prosecution history thereof and advise the USPTO that the claims in this application may be broader than any claim in the parent application(s).
The present invention relates to an apparatus and a method of encoding/decoding a video, and more particularly, to a video encoding/decoding method and apparatus which displays a block having a quantization/dequantization differential value based on a quadtree with respect to coding units (CUs) in a largest coding unit (LCU) and adaptively predicts/decodes a quantization/dequantization parameter value using context information on blocks neighboring to a block to encode/decode.
High Efficiency Video Coding (HEVC) encodes/decodes an input picture by a coding unit (CU). A CU with a largest size in a frame is referred to as a largest coding unit (LCU), which is split into a plurality of CUs based on quadtree splitting information for encoding/decoding. In HEVC, one quantization parameter value is allocated to an LCU, and a quantization parameter value for a current LCU to encode is predicted from a previous LCU according to raster scan order.
In H.265/AVC, encoding/decoding is carried out on a macroblock unit, and a quantization/dequantization value is allocated to each macroblock. A quantization parameter value allocated to each macroblock is predicted from a quantization parameter value of a macroblock located on the left in a frame. A differential value which is generated after prediction of the quantization parameter value is written in a macroblock to encode, thereby conducting encoding. A decoder decodes the quantization parameter value by adding the quantization parameter differential value decoded in entropy decoding and the quantization parameter value of the left macroblock.
However, when a large LCU is allocated as compared with a size of an input picture, a bit rate may not be effectively controlled using a quantization parameter value stored by an LCU. Further, when a quantization parameter value is allocated by a CU, subjective deterioration in image quality may occur due to a difference in quantization parameter value from neighboring CUs. Thus, there is a need for a method of allocating quantization parameter values to various block sizes ranging from a CU to an LCU depending on an input picture and of optimally predicting a quantization parameter using context information on blocks neighboring to a block to encode.
An aspect of the present invention is to provide a method and an apparatus for encoding/decoding quadtree-based quantization/dequantization parameter values based on splitting information on a CU. The quantization/dequantization parameter encoding/decoding method and apparatus are capable of predicting quantization/dequantization parameter values in an effective manner using context information on neighboring blocks.
Technical problems of the present invention are not limited to those mentioned above, and those skilled in the art may understand additional problems not stated herein based on the following description.
An embodiment of the present invention provides a video encoding apparatus which includes a quantization differential value storing block unit determination module to determine a block unit to store a quantization parameter (QP) differential value in splitting a largest coding unit (LCU) of a picture into a plurality of CUs in a quadtree or encoding into a single CU, a quantization module to perform quantization using a quantization value allocated to each block unit, a quantization prediction block determination module to adaptively determine a prediction block using context information on a neighboring block so as to predict a quantization value used by a block to encode, a QP differential value generation module to generate a quantization differential value of the block to encode using a QP of the prediction block determined based on the context information, and a QP storage module to store splitting information on a quantization differential value storage block unit and QP differential values of a corresponding block.
Another embodiment of the present invention provides a video decoding apparatus which includes a dequantization parameter differential value block splitting flag deriving module to decode information on a block having a dequantization parameter differential value in an LCU, a dequantization differential value storage block unit determination module to determine the block having the dequantization parameter differential value in the LCU using a decoded dequantization parameter differential value block splitting flag, a dequantization parameter differential value deriving module to decode a dequantization parameter differential value according to the dequantization parameter differential value block splittting flag, a dequantization parameter value prediction block determination module to determine a block used for prediction based on context information on a neighboring block so as to decode a dequantiation parameter value of a block to decode, a dequantization parameter value deriving module to decode the dequantization parameter value used for dequantization, and a dequantization module to perform dequantization using the decoded dequantization parameter value.
Quadtree-based adaptive quantization/dequantization parameter encoding and decoding methods and apparatuses according to exemplary embodiments of the present invention enable allocation of different levels of quantization parameter differential values when a block is split in a quadtree. Such allocation of different levels of quantization parameter differential values allows accurate adjustment of bit rate as compared with allocation of a single quantization parameter value in an LCU. Further, in predicting/decoding a quantization/dequantization parameter value for each block unit, a prediction direction may be adaptively determined using not only quadtree-based zigzag scanning but also context information on neighboring blocks, thereby resolving subjective deterioration in image quality which may occur due to a substantial difference in quantization value from neighboring blocks.
Hereinafter, quadtree-based adaptive quantization/dequantization parameter encoding and decoding apparatuses according to an exemplary embodiments of the present invention will be described with reference to the accompanying drawings.
1 FIG.A illustrates a method and an apparatus for adaptively allocating a quantization parameter (QP) value to a block having a quadtree structure and encoding the QP value in a video encoding apparatus according a first exemplary embodiment of the present invention.
1 FIG.A 100 101 102 103 104 Referring to, a quadtree-based adaptive quantization parameter encoding method and apparatus include a quantization differential value storage block unit determination module, a quantization module, a QP value prediction block determination module, a QP differential value generation module, and a QP storage module.
100 The quantization differential value storage block unit determination modulemay determine a block unit to record a quantization differential value for each coding unit (CU) or a group of a plurality of CUs based on information on splitting of a largest coding unit (LCU) of a picture into CUs. Information on a block storing the quantization differential value may have a quadtree structure.
101 The quantization moduleperforms quantization on an input block using a QP value allocated to the block.
102 The QP value prediction block determination moduledetermines a prediction block to be used for prediction of a quantization value allocated to each CU or arbitrary CU using context information on a CU neighboring to the CU so as to effectively encode the quantization value. The context information may include a size of the block, a location of the block, and a prediction mode for the block.
103 102 The QP differential value generation modulegenerates a QP differential value by subtracting a quantization value of the current block from a QP value of the quantization value prediction block determined by the QP value prediction block determination module.
104 The QP storage moduleserves to entropy-encode a sequence parameter set, flag information on application/non-application by a slice unit, flag information used to indicate splitting information on a block including a quantization differential value, and a QP differential value for quadtree-based adaptive QP encoding.
1 FIG.B illustrates a quadtree-based adaptive dequantization parameter decoding method and apparatus in a video decoding apparatus according the first exemplary embodiment of the present invention.
1 FIG.B 120 121 122 123 124 125 Referring to, the quadtree-based adaptive dequantization parameter decoding method and apparatus includes a dequantization parameter differential value block splitting flag deriving module, a dequantization differential value storage block unit determination module, a dequantization parameter differential value deriving module, a dequantization parameter value prediction block determination module, a dequantization parameter value deriving module, and a dequantization module.
120 The dequantization parameter differential value block partition flag deriving moduledecodes a sequence parameter set and a block partition flag for a block having a dequantization parameter differential value by an LCU in slice data.
121 The dequantization differential value storage block unit determination moduledetermines blocks to record dequantization parameter differential values using the decoded dequantization parameter differential value block splitting flag and a CU splitting flag. Information on a block to record a quantization differential value may have a quadtree structure.
122 121 The dequantization parameter differential value deriving modulederives a dequantization parameter differential value of each block determined by the dequantization differential value storage block unit determination module.
123 The dequantization parameter value prediction block determination modulederives determines a block for adaptive reference using context information on a neighboring block in dequantization. The context information may include a size of the block, a location of the block, and a prediction mode for the block.
124 125 123 122 The dequantization parameter value deriving modulederives a dequantization parameter value used for the dequantization moduleby adding a dequantization parameter value of a prediction block derived by the dequantization parameter value prediction block determination moduleand a dequantization parameter differential value derived by the dequantization parameter differential value deriving module.
125 124 The dequantization moduleperforms dequantization on the input block using the parameter derived by the dequantization parameter value deriving module.
2 FIG. illustrates a configuration of the video decoding apparatus according to the first exemplary embodiment of the present invention.
2 FIG. 200 210 220 230 240 250 260 Referring to, the vide decoding apparatus includes an entropy-decoding module, a quadtree-based dequantization parameter deriving module, a rearrangement module, a dequantization module, an inverse discrete cosine transform encoding module, an intra/inter prediction module, and a filtering module.
200 120 121 122 The entropy-decoding moduleincludes the dequantization parameter differential value block splitting flag deriving moduleto derive a block partition flag used for quadtree-based adaptive dequantization, the dequantization differential value storage block unit determination moduleto determine a block storing a QP from the derived block splitting flag, and the dequantization parameter differential value deriving moduleto decode a dequantization parameter differential value stored in a corresponding block.
210 123 124 The quadtree-based dequantization parameter deriving moduleincludes the dequantization parameter value prediction block determination moduleto determine a prediction block for reference in decoding a dequantization parameter and the dequantization parameter value deriving moduleto derive a dequantization parameter by adding a dequantization parameter of the prediction block and the derived dequantization parameter differential value.
3 FIG. illustrates a context for controlling a quadtree-based quantization differential value stored in a sequence parameter set according to the first exemplary embodiment of the present invention.
300 When cu_qp_delta_enabled_flaghas a value of 1 in the sequence parameter set, quantization/dequantization parameter differential values of various quadtree blocks ranging from CUs of minimum size to CUs of maximum size in all slices in a sequence may be controlled.
4 FIG.A illustrates variables set to initial values in slice data according to the first exemplary embodiment of the present invention.
When a slice is split in a quadtree and encoded/decoded, the slice is first split into LCUs that are a quadtree of maximum size, which are then encoded/decoded in sequential scanning order. In encoding/decoding each LCU, an LCU may be further split into a plurality of CUs in a quadtree, and such splitting may be performed until CUs are partitioned into CUs of minimum size.
4 FIG.A 400 In, is CuQpDeltaCodedis a variable for controlling a quantization/dequantization parameter differential value which may be stored in each CU when an arbitrary CU is split into N CUs. This variable is always initialized to 0 before each LCU in the slice is encoded/decoded.
4 FIG.A 401 In, coding_treeis a function of performing encoding/decoding on one LCU in the slice. A fourth factor in this function is a flag indicating whether a quantization/dequantization parameter differential value exists in a corresponding CU, and always imports a value of 1 before each LCU in the slice is encoded/decoded since at least one quantization/dequantization parameter differential value is stored in the LCU.
4 FIG.B illustrates a context of dequantization parameter differential value block splitting stored in a coding tree block according to the first exemplary embodiment of the present invention.
420 A coding tree block expresses a context of a CU. A CU having a 2N×2N size may be split into four CUs having an N×N size and encoded/decoded according to split_coding_unit_flag. Alternatively, the CU having a 2N×2N size may not be split into CUs of smaller size any more but be encoded/decoded as it is.
421 420 422 300 421 422 421 The current CU receives an input of a flag indicating whether a quantization/dequantization parameter differential value exists in a higher CU, that is, cu_qp_delta_exist_flag. When the current CU having a 2N×2N size is further split into CUs having an N×N size according to a value of split_coding_unit_flag, split_qp_delta_flagis additionally encoded/decoded. Information on such additional splitting is encoded/decoded only when cu_qp_delta_enable_flagstored in the sequence parameter set and cu_qp_delta_exist_flaginput from the higher CU have a value of 1. A value of split_qp_delta_flagis encoded/decoded only when the current CU having a 2N×2N size is split into the CU having an N×N size, and is input as the value of cu_qp_delta_exist_flagwhen the lower CUs having an N×N size are encoded/decoded.
422 422 400 When the value of split_qp_detla_flagis 0, the current CU having a 2N×2N size is partitioned into the CUs having an N×N size but a block storing a quantization/dequantization parameter differential value is not split any more from a 2N×2N size to an N×N size. When the value of split_qp_detla_flagis 0, a value of IsCuQpDeltaCodedis additionally initialized to 0, so that a quantization/dequantization parameter differential value is stored only in a first N×N CU when the current CU is split into the CUs having N×N size.
5 FIG. illustrates a context of a quantization/dequantization parameter differential value stored by a CU and conditions in which the differential value exists according to the first exemplary embodiment of the present invention.
500 421 501 300 500 421 300 501 When a CU is not in a skip mode, a quantization/dequantization parameter differential value may be recorded in the CU. When a value of cu_qp_delta_exist_flagandis 1, a quantization/dequantization parameter differential value exists in the current CU, in which case a value of cu_qp_deltamay be stored by a CU according to a value of cu_qp_delta_enabled_flagstored in the sequence parameter set. For example, when the value of cu_qp_delta_exist_flagandis 1 and the value of cu_qp_delta_enabed_flagis 0, cu_qp_deltais not stored.
422 500 501 501 501 501 Alternatively, when a 2N×2N CU is split into four N×N CUs in a quadtree, only one quantization/dequantization parameter differential value may be stored. In this case, the quantization/dequantization parameter differential value is stored in a first CU among the four CUs, while the other three CUs do not record the quantization/dequantization parameter differential value. Here, since the value of split_qp_delta_flagencoded/decoded in the 2N×2N CU is 0, the value of cu_qp_delta_exist_flaginput to the N×N CU is 0. Thus, although no cu_qp_deltabasically exists in the split four N×N CUs, a value of cu_qp_deltamay be stored in a first N×N CU using a variable NIsCuQpDeltaCoded. The value of cu_qp_deltais not stored in the other three CUs since the first CU changes the value of IsCuQpDeltaCoded to 1 after decoding cu_qp_delta.
501 300 300 Even in this case, the value of cu_qp_deltamay be stored only when the value of cu_qp_delta_enabled_flagstored in the sequence parameter set is 1, simultaneously checking the value of cu_qp_delta_enabled_flag.
6 FIG.A 100 121 illustrates operations of the quantization differential value storage block unit determination moduleand the dequantization differential value storage block unit determination moduleaccording to the first exemplary embodiment of the present invention.
422 6 FIG.A A 2N×2N LCU to encode/decode may be split into four N×N CUs, each of which may be further partitioned to be processed. Even though the 2N×2N LCU is partitioned into a plurality of CUs to be encoded/decoded, when split_qp_delta_flaghas a value of 0 as shown in, a quantization/dequantization parameter differential value is stored in a first CU of the LCU. A quantization/dequantization parameter value reconstructed in the first CU may be used for the other CUs of the LCU.
6 FIG.B 100 121 illustrates operations of the quantization differential value storage block unit determination moduleand the dequantization differential value storage block unit determination moduleaccording to a second exemplary embodiment of the present invention.
422 631 422 631 422 631 A 2N×2N LCU to encode/decode is split into four CUs in a first stage, a second CU of which is further split. In this instance, when a flag indicating splitting into the CUs is encoded/decoded in the first stage, a splitting flag with respect to a quantization/dequantization parameter value, which is split_qp_delta_flagand, is additionally encoded/decoded. When a value of split_qp_delta_flagandis 1, all four split CUs have a quantization/dequantization parameter differential value, and thus the splitting flag with respect to the quantization/dequantization parameter value, split_qp_delta_flagand, is additionally encoded/decoded.
422 631 Even in this case, although the second N×N CU is split into CUs until a third stage, the block splitting flag for quantization/dequantization, split_qp_delta_flagand, is 0, and thus a single quantization/dequantization parameter value is allocated for a plurality of CUs.
6 FIG.C 100 121 illustrates operations of the quantization differential value storage block unit determination moduleand the dequantization differential value storage block unit determination moduleaccording to a third exemplary embodiment of the present invention.
6 FIG.C 6 FIG.C 422 661 shows that a 2N×2N LCU to encode/decode is split into four CUs in a first stage, a second CU of which is further split. Even when the second CU is split into four CUs, each of which is further partitioned into four CUs, a size of a block to record a quantization/dequantization parameter differential value may be determined using a value of split_qp_delta_flagand. In, a CU is split to regions with up to three depth informations, while the quantization/dequantization parameter differential value has relatively up to two depth informations.
7 FIG.A 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to the first exemplary embodiment of the present invention.
100 An encoder allocates a QP value to CU blocks determined by the quantization differential value storage block unit determination moduleand employs a previously used QP value for remaining blocks as it is. Here, the QP value is predicted from a QP value for previous blocks, and only a QP different value is encoded.
122 A decoder decodes a dequantization parameter differential value in the dequantization parameter differential value deriving moduleand derives a dequantization parameter value by adding the differential value with a dequantization parameter value of a block used for prediction.
102 123 The QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleserve to determine a neighboring block for reference when the encoder and the decoder predict a QP value of a current block.
7 FIG.A 720 712 702 710 711 712 720 700 701 702 720 In, when a CUto encode/decode is allocated a quantization/dequantization parameter value, the parameter value is predicted using an average value, minimum value or maximum value of QP values of CUsandor Lc and Tc having a largest block size located on each boundary among CUs,andor La, Lb and Lc located on a left boundary of the current CUand CUs,andor Ta, Tb and Tc located on an upper boundary of the current CU.
7 FIG.B 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to an alternative example of the first exemplary embodiment of the present invention.
750 750 750 When an LCU is split into a plurality of CUs to be encoded/decoded, a QP value of a current CUto encode/decode is predicted using an average value, minimum value or maximum value of QP values of two CUs having a largest block size located on a left boundary and an upper boundary of the current CUamong CUs neighboring to the current CU.
8 FIG.A 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to a second exemplary embodiment of the present invention.
840 840 820 840 800 When a QP value of a current CUto encode/decode is predicted, a CU having a largest block size among CUs neighboring to the current CU is used for reference. In this case, when there are a plurality of CUs having a largest block size, an average value, minimum value or maximum value of QP values of CUs is used to predict the QP value of the current CU, the CUs including a top CUor La selected among CUs on a left boundary of the current CUand a leftmost CUor Ta preferentially used for reference among CUs on an upper boundary thereof.
8 FIG.B 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to an alternative example of the second exemplary embodiment of the present invention.
890 890 890 870 890 850 890 When an LCU is split into a plurality of CUs to be encoded/decoded, a QP value of a current CUto encode/decode is predicted using CUs having a largest block size for reference among CUs neighboring to the current CU. Here, when there are a plurality of CUs having a largest block size on a left boundary of the current CU, a top CUor La is used as a reference block. Likewise, when there are a plurality of CUs having a largest block size on an upper boundary of the current CU, a leftmost CUor Ta is used as a reference block. When the reference blocks on the left and top boundaries are determined, the QP value of the current CUto encode/decode is predicted using an average value, minimum value or maximum value of QP values of these two CUs.
9 FIG.A 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to a third exemplary embodiment of the present invention.
920 900 901 902 910 911 912 920 When a QP value of a current CUto encode/decode is predicted, all possible reference CUs,,,,andor Ta, Tb, Tc, La, Lb and Lc neighboring to the current CU are selected as reference blocks. The encoder predicts the QP value of the current CUto encode using an average value, minimum value or maximum value of QP values of all possible reference CUs and encodes a QP differential value.
920 The decoder decodes a dequantization parameter value of the CUto decode by adding the decoded QP differential value with an average value, minimum value or maximum value of dequantization parameter values of all possible reference CUs.
9 FIG.B 102 123 illustrates the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to an alternative example of the third exemplary embodiment of the present invention.
950 950 930 931 932 940 941 942 When a CUto encode/decode is located in an LCU, a QP value of the CUis predicted using all possible reference CUs,,,,andor La, Lb, Lc, Ta, Tb and Tc neighboring to the current CU.
950 The encoder predicts the QP value of the current CUto encode using an average value, minimum value or maximum value of QP values of all adjacent possible reference CUs and encodes a QP differential value therebetween only.
950 The decoder decodes a dequantization parameter value of the CUto decode by adding the decoded QP differential value with an average value, minimum value or maximum value of dequantization parameter values of all adjacent possible reference CUs.
10 FIG.A 102 123 illustrates an operation in the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to a fourth exemplary embodiment of the present invention.
1020 1012 1010 1011 1012 1020 When a CUlocated on a boundary between LCUs is encoded/decoded, a CUor Lc having a largest block size among possible reference CUs,andor La, Lb and Lc located on a left boundary of the current CUis used for reference. Here, when there are one or more CUs having a largest block size among the possible reference CUs on the left boundary, a top CU is selected as a reference block.
1020 1012 The encoder predicts a QP value of the CUto encode using a QP value of the CUor Lc selected on the left boundary as the reference block and encodes a QP differential value therebetween.
1020 1012 1020 The decoder decodes a dequantization parameter differential value of the CUto decode and adds the differential value with a dequantization parameter value of the CUor Lc, thereby decoding a dequantization parameter value of the CU.
10 FIG.B 102 123 illustrates an operation in the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to an alternative example of the fourth exemplary embodiment of the present invention.
1030 1030 1031 1032 1050 1030 When an LCU is split into a plurality of CUs to be encoded/decoded, a CUor La having a largest block size among CUs,andor La, Lb and Lc located on a left boundary of a CUto encode/decode is used for reference. Here, when there are one or more CUs having a largest block size among the CUs on the left boundary, a top CUor La is selected as a reference block.
1050 1030 The encoder predicts a QP value of the CUto encode using a QP value of the CUor La selected on the left boundary as the reference block and encodes a QP differential value therebetween.
1050 1030 1050 The decoder decodes a dequantization parameter differential value of the CUto decode and adds the differential value with a dequantization parameter value of the CUor La, thereby decoding a dequantization parameter value of the CU.
11 FIG.A 102 123 illustrates an operation in the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to a fifth exemplary embodiment of the present invention.
1120 1110 1111 1112 1220 When a CUlocated on a boundary between LCUs is encoded/decoded, all possible reference CUs,andor La, Lb and Lc located on a left boundary of the current CUare used as reference blocks.
1120 1110 1111 1112 The encoder predicts a QP value of the CUto encode using an average value, minimum value or maximum value of QP values of all possible reference CUs,andor La, Lb and Lc located on the left boundary and encodes a QP differential value.
1120 1120 1110 1111 1112 The decoder decodes a dequantization parameter differential value of the CUto decode and decodes a dequantization parameter value of the CUto decode by adding the differential value with an average value, minimum value or maximum value of dequantization parameter values of all possible reference CUs,andor La, Lb and Lc.
11 FIG.B 102 123 illustrates an operation in the QP value prediction block determination moduleand the dequantization parameter value prediction block determination moduleaccording to an alternative example of the fifth exemplary embodiment of the present invention.
1150 When an LCU is split into a plurality of CUs to be encoded/decoded, all possible reference CUs located on a left boundary of a CUto encode/decode are used as reference blocks.
1150 The encoder predicts a QP value of the CUto encode using an average value, minimum value or maximum value of QP values of all possible reference CUs located on the left boundary and encodes a QP differential value.
1150 1150 The decoder decodes a dequantization parameter differential value of the CUto decode and decodes a dequantization parameter value of the CUto decode by adding the differential value with an average value, minimum value or maximum value of dequantization parameter values of all possible reference CUs located on the left boundary.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
October 20, 2022
August 25, 2026
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.