CN101039432A - Method and apparatus for realizing self-adaptive quantization in coding process - Google Patents
Method and apparatus for realizing self-adaptive quantization in coding process Download PDFInfo
- Publication number
- CN101039432A CN101039432A CN 200710087418 CN200710087418A CN101039432A CN 101039432 A CN101039432 A CN 101039432A CN 200710087418 CN200710087418 CN 200710087418 CN 200710087418 A CN200710087418 A CN 200710087418A CN 101039432 A CN101039432 A CN 101039432A
- Authority
- CN
- China
- Prior art keywords
- block
- quantitative mode
- mode
- information
- decision
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Disclosed is a method and device for realizing self-adaption quantization in encoding process. The mainly adopted implemental schemes comprise: firstly, based on the content of the encoding image, selecting a coincidental quantizing mode; secondly, in the decision-making process of the quantizing mode, adopting the encoding type, the motion vector, etc. of two adjacent blocks as the decision-making gist for the quantizing mode selection of the current block; finally, decision-making of quantizing mode also considering about the different sensitivity of human visual capability to different image content. Thereby, the present invention can make full use of human visual capability and can realize the great improvement on subjective quality of the encoding images in same encoding rate.
Description
Technical field
The present invention relates to the coding techniques field, relate in particular to a kind of technology that is used for the quantitative mode decision-making of cataloged procedures such as image.
Background technology
In image encoding and decoding technique, carry out encoding process again after image need being divided into encoding block usually at the coding of image.To obtain corresponding block conversion coefficient for the image that is divided into encoding block through the orthogonal transform processing, described block conversion coefficient is again through quantizing to round calculating, and entropy coding just can obtain the code stream behind the compressed encoding of this image correspondence at last, thereby realized the encoding operation at image.
In the framework of video compression coding, for intraframe coding, the data of quantification are the image block transform coefficient values, and the data that quantize for interframe encode are residual error coefficients.Because the content information of image all is stored in conversion coefficient or the residual error coefficient, therefore, in cataloged procedure, can quantize to realize control by control at the picture quality behind the coding at image.
Image of the present invention comprises still image, a width of cloth of active images, the residual error image of adjacent two images of active images, target image of any images process computing gained or the like in the active images.
In coding, normally realize for the quantization operation of conversion coefficient by quantization matrix, for example can adopt following formula to quantize:
Wherein, Coe (i, j) for image block through the after the conversion (i, the j) value of position pixel abbreviate conversion coefficient as, QM (i j) be quantization matrix, Q (i j) is the transform coefficient values after quantizing to round, abbreviate as quantize after coefficient value, [] is for rounding calculating.
Because the details of the image of different content is represented the frequency of different images, and therefore human eye, at the image of different content, should adopt the different quantization methods that meets the human eye characteristic to the subjective sensation difference of image different piece.
At present, in JPEG (associating picture experts group), MPEG1 (MPEG, motion picture expert group), MPEG2 is in the picture coding standards such as MPEG4, all adopt fixing quantization matrix to realize for the quantification of coded image, wherein, be that quantization matrix is placed on image head in the jpeg picture coding standard, in picture coding standards such as MPEG1, MPEG2, MPEG4, then be that quantization matrix is placed in the sequence head; Therefore, for sequence image, the quantization matrix of MPEG picture coding standard is that each sequence has one, promptly adopts same fixing quantization matrix to realize quantification treatment at image at same sequence.
Because human eye evaluation to picture quality when watching image is to estimate according to the image subjective quality that human eye is perceived.Thereby, have only and adopt the quantization method of the visual characteristic that meets human eye that image is carried out quantification treatment, could obtain subjective picture quality preferably.Promptly at an image sequence, the quantization matrix that needs to select to be fit to carries out quantification treatment to it, like this, can obtain gratifying subjective picture quality.
Yet, because the picture material in a sequence is not identical, but have bigger variation equally, promptly the details of image has nothing in common with each other in same image sequence, therefore, if all adopt same quantization matrix to realize quantification treatment, obviously can't reach the image subjective quality behind the best quantization encoding for whole sequence.
Summary of the invention
The embodiment of the invention provides a kind of method and device of realizing adaptive quantizing in cataloged procedure, with under identical encoder bit rate condition, improves the image subjective quality after encoding.
The embodiment of the invention provides a kind of method of adaptive quantizing that realizes in cataloged procedure to comprise:
Obtain the parameter information of the adjacent block of current block;
Determine the quantitative mode of current block according to the parameter information of described adjacent block;
Utilize definite quantitative mode that the conversion coefficient of current block is quantized.
The embodiment of the invention provides a kind of device of realizing adaptive quantizing in cataloged procedure, comprising:
Adjacent piece parameter information acquiring unit is used to obtain the parameter information of the adjacent block of current block;
The quantitative mode decision package, the parameter information that is used for the adjacent block that obtains according to described adjacent piece parameter information acquiring unit is determined the quantitative mode of current block;
The quantification treatment unit is used to utilize the definite quantitative mode of described quantitative mode decision package that the conversion coefficient of current block is quantized.
The technical scheme that is provided by the embodiment of the invention as can be seen, because the embodiment of the invention is the visual characteristic realization according to human eye, the quantization matrix of promptly selecting to be fit in cataloged procedure carries out quantification treatment to it, therefore, when coding, image sequence is quantized to be adapted to well the content characteristic of image, thereby under identical encoder bit rate condition, can make the subjective quality of the image after encoding be greatly improved.
Description of drawings
Fig. 1 is the layout structure schematic diagram of current block and adjacent block;
The structural representation of the device that Fig. 2 provides for the embodiment of the invention.
Embodiment
The quantitative mode that the embodiment of the invention is used includes but not limited to forms such as quantization matrix, quantization parameter and quantization step, can realize coding-decoding operation to image by the quantitative mode of determining.
Specifically be to determine the corresponding quantization pattern in the embodiment of the invention, thereby make each zone in the image, can obtain optimal quantization pattern within the specific limits according to image information; Wherein, described image information includes but not limited to image content information, image coding information.Therefore, the quantitative mode that utilizes the embodiment of the invention to provide quantizes to calculate to image-region, can be under identical code check condition, and the subjective quality of order coding back image obtains bigger raising.
The method of the realization adaptive quantizing that the embodiment of the invention provides, promptly the quantitative mode decision-making technique comprises following several concrete implementation:
Mode one
Can whether adopt different quantitative modes according to the piecemeal pattern information decision-making current block of the adjacent block of current block, and when determining to adopt different quantitative modes, determine the concrete quality that needs the quantitative mode of employing, and then need to determine the quantitative mode of selection; Described block size can be M * N, M, N=2,4,8,16 or other size; Described piecemeal pattern information includes but not limited to: at least one item in the information such as piecemeal mode profile of branch block size, the direction character that divides block type, piecemeal, adjacent block;
In this mode, specifically can also be according to different frame types, in frame, whether adopt different quantitative modes according to the piecemeal pattern information decision-making current block of the adjacent block of current block with interframe; Described frame type includes but not limited to any frame forms that contain the piecemeal pattern information such as I frame, P frame, B frame.
Mode two
Whether the motion vector information decision-making current block according to the adjacent block of current block adopts different quantitative modes, and when determining to adopt different quantitative modes, specifically need to determine the quality of the quantitative mode that adopts, and then need to determine the quantitative mode of selection; Described motion vector information includes but not limited to the motion vector distribution of motion vector size, motion vector direction, adjacent block etc.
The specific implementation of this mode two can also be according to different frame types, and whether the adjacent block motion vector information of current block decision-making current block adopts different quantitative modes; Wherein, described frame type includes but not limited to any frame forms that contain motion vector information such as P frame, B frame.
Mode three
According to the quantitative mode quality of current block and the employing of adjacent block informix decision-making current block, and then definite corresponding quantization pattern.
The scheme that the embodiment of the invention provides a kind of quantification of content-adaptive to realize promptly adopts different quantitative modes by the part to different content in the image sequence, thereby can improve the subjective quality of coded image greatly under same code rate.
Specifically, the embodiment of the invention is to utilize the parameter information of the adjacent block in the cataloged procedure, and the quantitative mode of present encoding piece is made a strategic decision.The parameter information of adjacent block includes but not limited to: wherein at least one of the piecemeal pattern information of adjacent block, the information of forecasting of adjacent block and the motion vector information of adjacent block etc.These information of adjacent block all can reflect the image property of present encoding piece, therefore, can be used as the foundation of the quantitative mode of determining current block by at least one item in these information.
Take all factors into consideration human vision property, the corresponding criterion that need follow when current block being carried out the quantitative mode decision-making is including but not limited to following several:
(1) preservation that should try one's best for the detail section of image, the quantitative mode of corresponding employing should keep more high-frequency information, promptly adopts high-quality quantitative mode.
(2) for the movement edge part of image, should keep more high-frequency information, need to adopt high-quality quantitative mode equally.
(3) for the more violent zone of motion, the decline of suitable objective quality can not have very big influence to subjective quality, therefore, can adopt more coarse quantitative mode, promptly adopts low-quality quantitative mode.
In above-mentioned three kinds of situations,, can adopt the quantitative mode of different quality at the feature of image different piece; Wherein, described high-quality quantitative mode and low-quality quantitative mode can preestablish, simultaneously, the quantitative mode of more multistage different quality can also be set, for example, the quantization matrix of high, medium and low three kinds of quality is set, can selects the quantitative mode of using as each current block.The quantitative mode of multiple quantification quality specifically can at first be set in system, and the parameter information condition of the quantitative mode adjacent block of every kind of quantification of selection quality of setting correspondence or the parameter information condition of current block and adjacent block, afterwards, just can determine the quantitative mode that current block adopts according to the parameter information of adjacent block or the parameter information of current block and adjacent block.
For example, for the quantization matrix of different quality, can be divided into following a few class quantization pattern:
(1) quantitative mode 0:
Be the quantitative mode of acquiescence, promptly the quantization matrix that adopts when coded macroblocks is identical with the quantization matrix of current acquiescence;
(2) quantitative mode 1:
The pattern that keeps details, the coded image after promptly the quantization matrix that adopts when coded macroblocks should make quantification keeps image detail, and this quantitative mode is a kind of high-quality quantitative mode;
(3) quantitative mode 2:
Non-details quantitative mode, promptly current coding macro block is not an image detail, and the quantization encoding image need not to keep image detail, and this quantitative mode is a kind of low-quality quantitative mode.
Afterwards, set to select the parameter information condition of different quantitative modes again, like this, the just quantitative mode that can select current block to adopt according to the parameter information condition in coding side and decoding end.
That is to say, in the actual coding process, utilize the parameter information of adjacent block, by predefined criterion, the quantitative mode of current block is made a strategic decision, the quantitative mode of different quality of parameter information correspondence that promptly preestablishes adjacent block is as decision rule.The criterion of setting comprises: when which type of the parameter information of adjacent block meets and require, determine that current block need adopt predetermined high-quality quantitative mode, when what situation the parameter information of adjacent block meet, determine that current block need adopt predetermined low-quality quantitative mode.Wherein said high-quality quantitative mode is meant and keeps more frequency information, the quantitative mode with higher eyefidelity; Described low-quality quantitative mode is meant loses some frequency informations, the quantitative mode that eyefidelity is low.
And, for different frame types, promptly need adopt different quantitative mode decision rules at intraframe coding and interframe encode.For example, when in-frame encoding picture, for the general more complicated of image detail partial mode information, if therefore the piecemeal pattern information of each adjacent macroblocks (being adjacent block) has nothing in common with each other, then current block need adopt more high-quality quantitative mode; When inter coded images,, can utilize the basis of the pattern information of the block size information of adjacent block and adjacent block as the quantitative mode decision-making for the predictive mode more complicated in image detail zone.
For ease of understanding to the embodiment of the invention, will be described embodiment of the invention specific implementation quantitative mode decision-making treatment process below, mainly comprise following several disposition:
Different frame types is taked different decision rule;
Different frame types, i.e. two types of intraframe coding and interframe encodes, decision rule can be that different, above-mentioned adjacent macroblocks defines as accompanying drawing 1, and in in-frame encoding picture, the intra prediction mode of adjacent block, branch block type can be used as the foundation of decision-making; In inter coded images, information such as inter-frame forecast mode of adjacent block (comprise block size, divide block type) and motion vector can be used as the foundation of decision-making.
In Fig. 1, current block is E, and the contiguous block of E is A, B, C, D; And the piece size of A, B, C, D, E can be identical or different.Wherein, the vertical direction contiguous block of current block E is A and D; The horizontal direction contiguous block of current block E is B and D or C and D.Described contiguous block can be direct neighbor piece or the indirect neighbor's piece of E; The direct neighbor piece of described E, i.e. the adjacent boundary piece of E; Indirect neighbor's piece of described E, the i.e. adjacent block of the adjacent boundary piece of E.
Decision-making mode to the different frame type describes below:
(1) quantitative mode of intraframe coding image is made a strategic decision;
Be that present frame is the intraframe coding image, as the I frame, then adopt the definite current block corresponding quantization pattern of quantitative mode decision-making in the frame, described Intra-coded blocks is meant the piece that adopts intra prediction mode.
Be specially in-frame encoding picture, image detail part predictive mode more complicated, the pattern information of adjacent block is generally all inequality, if therefore the predictive mode of adjacent macroblocks is inequality, illustrate that current macro may be the details area or the edges of regions of image, then adopts the quantitative mode of high-quality reservation details to current macro; For example, in Fig. 1, have only when the prediction mode information of adjacent macroblocks satisfies one of following condition, current macro selected the quantitative mode of low-quality non-details:
(1) quantitative mode of the contiguous block of the vertical direction of E employing is identical, and is identical with the quantitative mode that piece D adopts as piece A;
(2) quantitative mode of the contiguous block of the horizontal direction of E employing is identical, and is identical with the quantitative mode that piece D adopts as piece B (perhaps piece C);
Described predictive mode is identical to be meant that the predictive mode of two pieces is the same, and for example, two pieces are the vertical prediction pattern of employing, or are the horizontal forecast pattern.
(2) quantitative mode of interframe encode image is made a strategic decision;
Be that present frame is the interframe encode image,, then adopt the decision-making of interframe quantitative mode to determine current block corresponding quantization pattern as P frame, B frame.
Be specially inter coded images, the general more complicated of block mode information of its image detail part, if therefore the interframe encoding mode of adjacent macroblocks is inequality, current macro may be the edge in details area or zone, then current macro is adopted the high-quality quantitative mode that keeps details, for example, in Fig. 1, have only when the prediction mode information of adjacent macroblocks satisfies one of following condition, can select the low-quality quantitative mode of non-details to current macro:
(1) adjacent piece A, B, C, neither one Intra-coded blocks among the D;
(2) adjacent piece A, B, C, neither one small size pattern-coding piece among the D;
(3) contiguous block of vertical direction, as A and D all is large scale pattern-coding piece;
(4) contiguous block of horizontal direction, as B (perhaps C) and D all is large scale pattern-coding piece.
In addition, violent regional insensitive according to human vision property to moving, therefore, can adopt low-quality quantitative mode to these parts.
And, encoded in the adjacent block, each macro block all contains motion vector, each motion vector all contains motion vector size and motion vector direction, therefore, the motion vector of adjacent each macro block, the motion vector distribution of adjacent macroblocks can be used for estimating the kinetic characteristic in present encoding zone, therefore also can be used as the foundation of interframe encode quantitative mode decision-making.
Described coded block pattern includes but not limited to information such as the block size of encoding block and predictive mode; So-called piece size is meant the size of employed of the infra-frame prediction compensation of Intra-coded blocks or interframe encode block motion compensation, interframe encode is mostly supported the motion compensation of many sized blocks in present coding scheme, so-called small size mode block is meant the less piece of size of employing, as 4 * 4; So-called large scale mode block is meant larger-size of employing, as 16 * 16, and SKIP piece etc.
Can obtain the quantitative mode of current block by above-mentioned processing procedure, and then can quantize coefficient data.
To be example with standard H.264/AVC below, described invention be realized the process of quantitative mode decision-making describes.
Different frame types is taked different decision rule;
Described frame type comprises the I frame in H.264/AVC, P frame, B frame etc.Above-mentioned adjacent macroblocks definition is still as accompanying drawing 1: in the I frame, the adjacent macroblocks intra prediction mode can be used as the foundation of decision-making; In the P frame, inter-frame forecast mode, information such as motion vector can be used as the foundation of decision-making.The predictive mode of adjacent macroblocks in the described frame, in H.264/AVC, the encoding block to 4 * 4 is 9 kinds of prediction direction; Described inter-frame forecast mode in H.264/AVC, is supported the motion compensation technique of many sizes, and the piece size has 16 * 16,16 * 8,8 * 16, and relevant informations such as 8 * 8.....
In H.264/AVC, the piece of intra prediction mode can be arranged in the inter prediction image.Described small size mode block can refer to the pattern of 8 * 8 and 8 * 8 following sizes in H.264/AVC.Described large scale mode block can refer to the mode block of 8 * 8 above sizes in H.264/AVC.
To be that situation with the boundary block of the border direct neighbor of E is that example describes the embodiment of the invention below with the contiguous block of supposition E.
(1), then adopts quantitative mode decision-making in the frame if present frame is the I frame
To in-frame encoding picture, image detail part predictive mode more complicated, the pattern information of adjacent block is generally all inequality.If therefore the pattern of adjacent macroblocks is inequality, then adopt high-quality quantitative mode.In Fig. 1, have only when the pattern information of adjacent macroblocks satisfies one of following condition, just can select low-quality quantitative mode:
(1) block mode of vertical adjacent boundary is identical, and identical with the pattern of piece D as piece A, described pattern is identical, for example is horizontal direction predictive mode etc.;
(2) block mode of horizontal adjacent boundary is identical, and is identical with the pattern of piece D as piece B (or piece C), for example, is horizontal direction predictive mode etc.
With reference to previously described quantitative mode 0,1,2,, specifically can determine the quantitative mode that current block E adopts by following decision-making mode at Fig. 1:
(1) as if having one to be directionless pattern or average (DC) pattern among A, B, C, the D, then E is adopted quantitative mode 0, i.e. Mo Ren quantitative mode;
(2) if the block mode of vertical adjacent boundary is identical, macro block mode as A, D is identical, for example, be horizontal direction predictive mode etc., perhaps if the block mode of horizontal adjacent boundary is identical, identical as the macro block mode of B, D, perhaps the macro block mode of B, C is identical, then E is adopted quantitative mode 0, promptly give tacit consent to quantitative mode;
(3) as if A, B, discontented (1) and (2) condition that is enough to of C, D, then E is adopted quantitative mode 1, promptly keep the quantitative mode of details.
(2), adopt the decision-making of interframe quantitative mode if present frame is the P frame
In Fig. 1, have only when the pattern information of adjacent macroblocks satisfies one of following condition, just select lower quantitative mode:
(1) adjacent piece A, B, C, neither one Intra-coded blocks among the D;
(2) adjacent piece A, B, C, neither one small size mode block among the D;
(3) block mode of vertical adjacent boundary is identical, and is identical with the pattern of D as A, all is the large scale mode block;
(4) block mode of horizontal adjacent boundary is identical, and is identical with the pattern of D as B (or C), all is the large scale mode block.
Still describe and comprise and then in the P frame, current block E acquiescence is adopted quantitative mode 0,1,2 three kind quantitative mode 0, and comprise at the quantification decision-making mode of current block E with reference to the front:
(1) as if having one to be Intra-coded blocks among A, B, C, the D, then E is adopted quantitative mode 1, promptly keep detail modes;
(2) as if having one to be the small size piece among A, B, C, the D, as 4 * 4 block modes or 8 * 8 block modes, then E is adopted quantitative mode 1, promptly keep detail modes;
(3) if the piece of vertical adjacent boundary or/and the piece of horizontal adjacent boundary are the skip pattern, piece A and D as vertical adjacent boundary are the skip pattern, or/and piece B (or C) and the D of horizontal adjacent boundary is the skip pattern, the predictive mode that current block E is described also may be the skip pattern, the a certain macro block that skip mode declaration pattern specification current macro E has just in time encoded corresponding to former frame, but can't judge whether current macro E is image detail, therefore, E is adopted quantitative mode 0, promptly give tacit consent to quantitative mode; E is also adopted quantitative mode 1, promptly keep detail modes.
(4) if the piece of vertical adjacent boundary or the piece of horizontal adjacent boundary are the large scale block mode, as A, D is that 16 * 16 block modes or B, D are 16 * 16 block modes, perhaps B, C are 16 * 16 block modes, illustrate that current macro E is non-details area probably, then E is adopted quantitative mode 2, the quantitative mode of promptly non-reservation details;
(5) as if A, B, discontented (1)～(4) relation that is enough to of C, D,, promptly keep the quantitative mode of details then to the quantitative mode 1 of E employing.
Equally, if present frame is the B frame, then also adopt the decision-making of interframe quantitative mode at current block E, concrete decision-making mode is as follows:
(1) if the piece of vertical adjacent boundary or/and the piece of horizontal adjacent boundary are the skip pattern, as piece A, the D of vertical adjacent boundary is that piece B (or C), the D of skip pattern or/and horizontal adjacent boundary is the skip pattern, then E is adopted quantitative mode 0, i.e. Mo Ren quantitative mode;
(2) if the piece of vertical adjacent boundary or/and the piece of horizontal adjacent boundary are 16 * 16 patterns, as piece A, the D of vertical adjacent boundary is that piece B (or C), the D of 16 * 16 patterns or/and horizontal adjacent boundary is 16 * 16 patterns, then E is adopted quantitative mode 2, the quantitative mode of promptly non-reservation details;
(3) as if A, B, discontented (1) and (2) relation that is enough to of C, D, then E is adopted quantitative mode 1, promptly keep the quantitative mode of details.
The embodiment of the invention also provides a kind of device of realizing adaptive quantizing in cataloged procedure, is used for the quantitative mode according to the coding parameter information decision current block of contiguous block, and as shown in Figure 2, the specific implementation structure of this device comprises:
Adjacent piece parameter information acquiring unit is used to obtain the parameter information of the adjacent block of current block; And described adjacent piece parameter information acquiring unit is specifically as follows adjacent macroblocks coded message extractor etc.;
The quantitative mode decision package, the parameter information that is used for the adjacent block that obtains according to described adjacent piece parameter information acquiring unit is determined the quantitative mode of current block; And described quantitative mode decision package specifically can comprise the quantitative mode determining unit of intraframe coding quantitative mode decision-making device, interframe encode quantitative mode decision-making device and current block;
The quantification treatment unit is used to utilize the definite quantitative mode of described quantitative mode decision package that the conversion coefficient of current block is quantized; And described quantification treatment unit is specifically as follows quantizer or inverse DCT.
Described device can also comprise the frame type grader, and being used for provides the frame type parameter information with the quantitative mode decision package.
To the each several part that described quantitative mode decision making device specifically comprises be described respectively below:
Described frame type grader is used to obtain the type of the frame at present encoding piece place, as I frame, P frame or B frame etc.; And with coded quantization mode decision device and interframe encode quantitative mode decision-making device in the type information notification frame of the frame at present encoding piece place; Different frame types will influence the result of the selection and the quantitative mode decision-making of decision device.
Described adjacent macroblocks coded message extractor, the vicinity that is used to obtain the present encoding piece is the parameter information of encoding block, and described parameter information comprises at least one in the motion vector information of information of forecasting, contiguous block of piecemeal pattern information, the contiguous block of contiguous block and the quantitative mode information that contiguous block adopts; And with coded quantization mode decision device in the described parameter information notification frame and interframe encode quantitative mode decision-making device; So that according to the coding parameter information of different vicinity encoding block, the different decision device in the trade-off decision device.
Described intraframe coding quantitative mode decision-making device is used for the encoding block quantitative mode of making a strategic decision current according to the frame type and the intra coded information of the contiguous block that obtains; The intra coded information of described contiguous block comprises the intra prediction direction information or the quantitative mode information of block size information, piecemeal pattern information, piece; Typically, the quantitative mode decision-making device can be divided into contiguous block size decision device, contiguous block type decision device and/or contiguous block intra prediction mode (or direction) decision device etc. according to the difference of the intra coded information of the contiguous block of input in the frame.
Described interframe encode quantitative mode decision-making device is used for the encoding block quantitative mode of making a strategic decision current according to the frame type and the interframe encode information of the contiguous block that obtains; The interframe encode information of described contiguous block comprises motion vector direction, the size information of inter prediction directional information, the piece of block size information, piecemeal pattern information, piece, quantitative mode information of contiguous block etc.; Typically, interframe quantitative mode decision-making device can be divided into the motion vector decision device of contiguous block size decision device, contiguous block type decision device, contiguous block inter-frame forecast mode (or direction) decision device and/or contiguous block etc. according to the difference of interframe encode information of the contiguous block of input.
The quantitative mode determining unit of described current block, be used for determining with final selection the quantitative mode of current block according to the output of intraframe coding quantitative mode decision-making device or interframe encode quantitative mode decision-making device, be the class in some quantitative modes, be used for being input to quantizer, upgrade the parameter value of quantizer, with the quantification quality of control subsequent quantizatiion device; Described quantitative mode can be for multiple, and typical pattern comprises, keeps the quantitative mode of the quantitative mode of image detail, non-reservation image detail, the quantitative mode identical with the acquiescence quantification etc.
The operation principle of this quantitative mode decision making device is according to frame type and the contiguous coding parameter information of encoding block, to select suitable quantitative mode decision device.
(1) if in-frame encoding picture, then select quantitative mode decision-making device in the frame, again according to the intra coded information of contiguous block, select a kind of decision device in contiguous block size decision device, contiguous block type decision device, contiguous block intra prediction mode (or direction) decision device etc., and the output mode result of decision;
Wherein,
Contiguous block size decision device is used for carrying out the quantitative mode decision-making according to the contiguous block dimension information;
Contiguous block type decision device is used for carrying out the quantitative mode decision-making according to the contiguous block type information;
Contiguous block intra prediction mode decision device is used for carrying out the quantitative mode decision-making according to contiguous block intra prediction mode information.
(2) if inter coded images is then used intraframe coding quantitative mode decision-making device according to the type selecting of present encoding piece, or interframe encode quantitative mode decision-making device:
(21) if the present encoding piece uses intraframe coding, then use intraframe coding quantitative mode decision-making device, this moment, the decision-making device structure can be identical with (1), and still, the coded message of the contiguous block that input is used can be different with (1); When inter coded images was used intraframe coding quantitative mode decision-making device, the coding parameter information of contiguous block can comprise more form, can comprise different modes such as 8 * 8,16 * 8,8 * 16,16 * 16 as the size of contiguous block;
(22) if the present encoding piece uses interframe encode, then use interframe encode quantitative mode decision-making device; Specifically according to the interframe encode information of contiguous block, a kind of decision device in the motion vector decision device of selection contiguous block size decision device, contiguous block type decision device, contiguous block inter-frame forecast mode (or direction) decision device, contiguous block etc., the output mode result of decision;
Wherein,
Contiguous block size decision device is used for carrying out the quantitative mode decision-making according to the contiguous block dimension information;
Contiguous block type decision device is used for carrying out the quantitative mode decision-making according to the contiguous block type information;
Contiguous block inter-frame forecast mode decision device is used for carrying out the quantitative mode decision-making according to contiguous block inter-frame forecast mode information;
The motion vector decision device of contiguous block is used for carrying out the quantitative mode decision-making according to the motion vector information of contiguous block.
In sum, the quantitative mode decision scheme that the embodiment of the invention provides has the following advantages:
Use this quantitative mode decision-making technique can obtain adaptive quantizing based on macroblock size when picture coding, promptly each macro block can have independently quantitative mode;
And, in image block level cataloged procedure, can when reaching the macro-block-level self-adaptive quantification, in image level and macro-block level code stream, need not to transmit quantization matrix, making does not have extra code stream expense in the picture block level;
In addition, the embodiment of the invention is the visual characteristic realization according to human eye, like this, when coding, make the content characteristic that image sequence is quantized to be adapted to well image, thereby under identical encoder bit rate condition, the subjective quality of the image behind the coding is greatly improved.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.
Claims (16)
1, a kind of method that realizes adaptive quantizing in cataloged procedure is characterized in that, comprising:
Obtain the parameter information of the adjacent block of current block;
Determine the quantitative mode of current block according to the parameter information of described adjacent block;
Utilize definite quantitative mode that the conversion coefficient of current block is quantized.
2, method according to claim 1 is characterized in that, described parameter information comprises:
At least one item in the quantitative mode information that the motion vector information of the piecemeal pattern information of frame type information, adjacent block, the information of forecasting of adjacent block, adjacent block and adjacent block adopt.
3, method according to claim 2 is characterized in that, described piecemeal pattern information comprises following at least one:
The branch block size definite according to block size;
The branch block type definite according to the mode type of the piecemeal of encoding;
The direction character of the piecemeal of determining according to the piecemeal directional information of the mode type of the piecemeal of encoding;
Determine the piecemeal mode profile information of adjacent block according to the spatial distribution relationship characteristic of the piecemeal pattern information of the branch block mode of adjacent block in the image.
4, method according to claim 2 is characterized in that, described motion vector information comprises following at least one:
The motion vector distribution of motion vector size, motion vector direction, adjacent block.
5, according to each described method in the claim 1 to 4, it is characterized in that the step of the quantitative mode of described definite current block comprises:
In frame, determine the quantitative mode of current block according to the parameter information of described adjacent block with interframe.
6, the implementation method of the adaptive quantizing in the encoder according to claim 5 is characterized in that, the step of the quantitative mode of described definite current block specifically comprises:
In frame and interframe determine the quality of the quantitative mode that current block adopts and then definite corresponding quantitative pattern according to the piecemeal pattern information of the adjacent block of current block;
And/or,
In frame and interframe determine the quality of the quantitative mode that current block adopts and then definite corresponding quantitative pattern according to the motion vector information of the adjacent block of current block;
And/or,
In frame, determine the quantitative mode quality that current block adopts according to the piecemeal pattern information and the motion vector information of current block and adjacent block with interframe, and then definite corresponding quantitative pattern.
7, according to each described method in the claim 1 to 4, it is characterized in that described method also comprises: the mode that the selection quantitative mode of various frame type correspondences is set;
And the step of the quantitative mode of described definite current block comprises:
Mode according to the selection quantitative mode of in-frame encoding picture and inter coded images correspondence is that current block is selected the corresponding quantization pattern.
8, according to each described method in the claim 1 to 4, it is characterized in that the step of the quantitative mode of described definite current block comprises:
Independently determine to have the quantitative mode of corresponding quantization quality for each image blocking matching.
9, according to each described method in the claim 1 to 4, it is characterized in that described quantitative mode is: quantization matrix, quantization parameter or quantization step.
10, according to each described method in the claim 1 to 4, it is characterized in that described method also comprises:
The quantitative mode of multiple quantification quality is set, and the parameter information condition of the adjacent block of the quantitative mode of every kind of quantification of selection quality of setting correspondence, the current block of the quantitative mode of every kind of quantification of selection quality of perhaps, setting correspondence and the parameter information condition of adjacent block.
11, a kind of device of realizing adaptive quantizing in cataloged procedure is characterized in that, comprising:
Adjacent piece parameter information acquiring unit is used to obtain the parameter information of the adjacent block of current block;
The quantitative mode decision package, the parameter information that is used for the adjacent block that obtains according to described adjacent piece parameter information acquiring unit is determined the quantitative mode of current block;
The quantification treatment unit is used to utilize the definite quantitative mode of described quantitative mode decision package that the conversion coefficient of current block is quantized.
12, device according to claim 11, it is characterized in that, described adjacent piece parameter information acquiring unit is an adjacent macroblocks coded message extractor, the vicinity that is used to obtain the present encoding piece encoding block parameter information and notify the quantitative mode decision package, described parameter information comprises at least one in the motion vector information of information of forecasting, contiguous block of piecemeal pattern information, the contiguous block of contiguous block and the quantitative mode information that contiguous block adopts.
13, device according to claim 11 is characterized in that, described device also comprises the frame type grader, is used to obtain the type of the frame at present encoding piece place, and notice quantitative mode decision package.
According to claim 11,12 or 13 described devices, it is characterized in that 14, described quantitative mode decision package comprises:
Intraframe coding quantitative mode decision-making device is used for the encoding block quantitative mode of making a strategic decision current according to the frame type and the intra coded information of the contiguous block that obtains;
Interframe encode quantitative mode decision-making device is used for the encoding block quantitative mode of making a strategic decision current according to the frame type and the interframe encode information of the contiguous block that obtains;
The quantitative mode determining unit of current block is used for selecting to determine the quantitative mode of current block according to the output of intraframe coding quantitative mode decision-making device or interframe encode quantitative mode decision-making device, and exports to the quantification treatment unit.
15, device according to claim 14 is characterized in that, the quantitative mode decision-making device specifically comprises at least one in contiguous block size decision device, contiguous block type decision device and the contiguous block intra prediction mode decision device in the described frame, wherein,
Contiguous block size decision device is used for carrying out the quantitative mode decision-making according to the contiguous block dimension information;
Contiguous block type decision device is used for carrying out the quantitative mode decision-making according to the contiguous block type information;
Contiguous block intra prediction mode decision device is used for carrying out the quantitative mode decision-making according to contiguous block intra prediction mode information.
16, device according to claim 14, it is characterized in that, described interframe quantitative mode decision-making device specifically comprises at least one in the motion vector decision device of contiguous block size decision device, contiguous block type decision device, contiguous block inter-frame forecast mode decision device and contiguous block, wherein
Contiguous block size decision device is used for carrying out the quantitative mode decision-making according to the contiguous block dimension information;
Contiguous block type decision device is used for carrying out the quantitative mode decision-making according to the contiguous block type information;
Contiguous block inter-frame forecast mode decision device is used for carrying out the quantitative mode decision-making according to contiguous block inter-frame forecast mode information;
The motion vector decision device of contiguous block is used for carrying out the quantitative mode decision-making according to the motion vector information of contiguous block.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200710087418 CN101039432A (en) | 2006-03-16 | 2007-03-16 | Method and apparatus for realizing self-adaptive quantization in coding process |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200610064858 | 2006-03-16 | ||
| CN200610064858.X | 2006-03-16 | ||
| CN 200710087418 CN101039432A (en) | 2006-03-16 | 2007-03-16 | Method and apparatus for realizing self-adaptive quantization in coding process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101039432A true CN101039432A (en) | 2007-09-19 |
Family
ID=38890026
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200710087418 Pending CN101039432A (en) | 2006-03-16 | 2007-03-16 | Method and apparatus for realizing self-adaptive quantization in coding process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101039432A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009086766A1 (en) * | 2007-12-04 | 2009-07-16 | Huawei Technologies Co., Ltd. | A quantization mode and an encoding and decoding method, an encoder and a decoder and a codec system |
| CN102006472A (en) * | 2010-11-18 | 2011-04-06 | 无锡中星微电子有限公司 | Video bitrate control system and method thereof |
| CN102196257A (en) * | 2010-03-11 | 2011-09-21 | 中国科学院微电子研究所 | Method and device for prejudging macroblock coding mode |
| CN103118255A (en) * | 2013-01-25 | 2013-05-22 | 深圳广晟信源技术有限公司 | Self-adaptation quantifying method based on concave model and self-adaptation quantifying device based on concave model |
| CN103828369A (en) * | 2011-06-10 | 2014-05-28 | 茨特里克斯系统公司 | Macroblock-level adaptive quantization in quality-aware video optimization |
| CN103871080A (en) * | 2014-03-25 | 2014-06-18 | 上海交通大学 | Method for adaptively quantizing optical flow features on complex video monitoring scenes |
| CN103975589A (en) * | 2011-12-13 | 2014-08-06 | Jvc建伍株式会社 | Video coding device, video coding method, video coding program, video decoding device, video decoding method, and video decoding program |
| CN105681811A (en) * | 2010-07-15 | 2016-06-15 | 夏普株式会社 | Decoding device and coding device |
| CN106878738A (en) * | 2011-01-24 | 2017-06-20 | 索尼公司 | Method for encoding images and message processing device |
| CN107181951A (en) * | 2011-03-09 | 2017-09-19 | 日本电气株式会社 | Video decoding apparatus and video encoding/decoding method |
| CN107431806A (en) * | 2015-03-19 | 2017-12-01 | Lg 电子株式会社 | For handling the method and its equipment of vision signal |
| CN109257603A (en) * | 2018-10-26 | 2019-01-22 | 西安科锐盛创新科技有限公司 | Bandwidth reduction quantization and quantification method |
| CN109510984A (en) * | 2018-10-26 | 2019-03-22 | 西安科锐盛创新科技有限公司 | A kind of compressed encoding quantization method |
| CN109561306A (en) * | 2018-10-26 | 2019-04-02 | 西安科锐盛创新科技有限公司 | A kind of bandwidth reduction quantization method |
| CN114051140A (en) * | 2022-01-07 | 2022-02-15 | 腾讯科技（深圳）有限公司 | Video encoding method, video encoding device, computer equipment and storage medium |
| CN117111637A (en) * | 2023-09-26 | 2023-11-24 | 江苏星湖科技有限公司 | Method and device for controlling patrol of convergence and fusion terminal of power high-altitude sensing equipment |
-
2007
- 2007-03-16 CN CN 200710087418 patent/CN101039432A/en active Pending
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101453643B (en) * | 2007-12-04 | 2011-05-18 | 华为技术有限公司 | Quantitative mode, image encoding, decoding method, encoder, decoder and system |
| WO2009086766A1 (en) * | 2007-12-04 | 2009-07-16 | Huawei Technologies Co., Ltd. | A quantization mode and an encoding and decoding method, an encoder and a decoder and a codec system |
| CN102196257A (en) * | 2010-03-11 | 2011-09-21 | 中国科学院微电子研究所 | Method and device for prejudging macroblock coding mode |
| CN102196257B (en) * | 2010-03-11 | 2014-02-12 | 中国科学院微电子研究所 | Method and device for prejudging macroblock coding mode |
| CN105681811A (en) * | 2010-07-15 | 2016-06-15 | 夏普株式会社 | Decoding device and coding device |
| CN105681811B (en) * | 2010-07-15 | 2018-11-30 | 夏普株式会社 | Decoding apparatus, code device and coding/decoding method |
| CN102006472A (en) * | 2010-11-18 | 2011-04-06 | 无锡中星微电子有限公司 | Video bitrate control system and method thereof |
| CN106878738B (en) * | 2011-01-24 | 2019-11-08 | 索尼公司 | Image encoding method and information processing equipment |
| CN106878738A (en) * | 2011-01-24 | 2017-06-20 | 索尼公司 | Method for encoding images and message processing device |
| US10771790B2 (en) | 2011-03-09 | 2020-09-08 | Nec Corporation | Video decoding device and method using inverse quantization |
| US11483571B2 (en) | 2011-03-09 | 2022-10-25 | Nec Corporation | Video decoding device and method using inverse quantization |
| CN107181951A (en) * | 2011-03-09 | 2017-09-19 | 日本电气株式会社 | Video decoding apparatus and video encoding/decoding method |
| US11509909B2 (en) | 2011-03-09 | 2022-11-22 | Nec Corporation | Video decoding device and method using inverse quantization |
| US11496749B2 (en) | 2011-03-09 | 2022-11-08 | Nec Corporation | Video decoding device and method using inverse quantization |
| CN107181951B (en) * | 2011-03-09 | 2019-12-24 | 日本电气株式会社 | Video decoding apparatus and video decoding method |
| CN103828369B (en) * | 2011-06-10 | 2018-09-07 | 茨特里克斯系统公司 | Quality perceive it is video optimized in macro-block-level self-adaptive quantization |
| CN103828369A (en) * | 2011-06-10 | 2014-05-28 | 茨特里克斯系统公司 | Macroblock-level adaptive quantization in quality-aware video optimization |
| CN103975589A (en) * | 2011-12-13 | 2014-08-06 | Jvc建伍株式会社 | Video coding device, video coding method, video coding program, video decoding device, video decoding method, and video decoding program |
| CN103118255A (en) * | 2013-01-25 | 2013-05-22 | 深圳广晟信源技术有限公司 | Self-adaptation quantifying method based on concave model and self-adaptation quantifying device based on concave model |
| CN103871080A (en) * | 2014-03-25 | 2014-06-18 | 上海交通大学 | Method for adaptively quantizing optical flow features on complex video monitoring scenes |
| CN107431806A (en) * | 2015-03-19 | 2017-12-01 | Lg 电子株式会社 | For handling the method and its equipment of vision signal |
| CN107431806B (en) * | 2015-03-19 | 2020-07-31 | Lg 电子株式会社 | Method for processing video signal and apparatus therefor |
| CN109257603A (en) * | 2018-10-26 | 2019-01-22 | 西安科锐盛创新科技有限公司 | Bandwidth reduction quantization and quantification method |
| CN109257603B (en) * | 2018-10-26 | 2021-05-11 | 西安科锐盛创新科技有限公司 | Bandwidth compression quantization and inverse quantization method |
| CN109561306B (en) * | 2018-10-26 | 2021-08-06 | 上海九吾尊易信息科技有限公司 | Bandwidth compression quantization method |
| CN109510984B (en) * | 2018-10-26 | 2021-02-05 | 和宇健康科技股份有限公司 | Compression coding quantization method |
| CN109561306A (en) * | 2018-10-26 | 2019-04-02 | 西安科锐盛创新科技有限公司 | A kind of bandwidth reduction quantization method |
| CN109510984A (en) * | 2018-10-26 | 2019-03-22 | 西安科锐盛创新科技有限公司 | A kind of compressed encoding quantization method |
| CN114051140A (en) * | 2022-01-07 | 2022-02-15 | 腾讯科技（深圳）有限公司 | Video encoding method, video encoding device, computer equipment and storage medium |
| CN117111637A (en) * | 2023-09-26 | 2023-11-24 | 江苏星湖科技有限公司 | Method and device for controlling patrol of convergence and fusion terminal of power high-altitude sensing equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101039432A (en) | Method and apparatus for realizing self-adaptive quantization in coding process | |
| CN101317458B (en) | Method and device for implementing adaptive quantization in encoding course | |
| CN100345450C (en) | Deblocking filtering method and apparatus of video frequency or image | |
| CN1820512A (en) | Method, medium, and filter removing a blocking effect | |
| CN1767653A (en) | Bit rate control method | |
| CN1812580A (en) | Deblocking control method considering intra bl mode and multilayer video encoder/decoder using the same | |
| CN1929611A (en) | Methods and apparatus for video intraprediction encoding and decoding | |
| CN1666429A (en) | Method and system for selecting interpolation filter type in video coding | |
| CN1640149A (en) | Video data conversion device and video data conversion method | |
| CN1846446A (en) | Process and arrangement for encoding video pictures | |
| KR101433170B1 (en) | Method of encoding and decoding using the directional information of the prediction mode of the adjacent blocks for estimating the intra prediction mode, and apparatus thereof | |
| CN1921630A (en) | Apparatus and method for video encoding/decoding and recording medium | |
| CN1925619A (en) | Apparatus and method for video encoding and decoding and recording medium | |
| CN1794815A (en) | Optimisation of a quantisation matrix for image and video coding | |
| CN1929613A (en) | Apparatus and method for image encoding and decoding and recording medium | |
| CN1685734A (en) | Method and apparatus for controlling rate of video sequence, video encoding device | |
| CN1829328A (en) | Prediction image generating method and apparatus and image and video encoding/decoding method and apparatus using the same | |
| CN101069429A (en) | Method and apparatus for multi-layered video encoding and decoding | |
| CN1829329A (en) | Indicating regions within a picture | |
| CN101040532A (en) | Deblocking filter | |
| CN1659580A (en) | Spatial prediction based intra coding | |
| CN1738423A (en) | Method for controlling video code bit rate | |
| CN101043621A (en) | Self-adaptive interpolation process method and coding/decoding module | |
| CN1777283A (en) | Microblock based video signal coding/decoding method | |
| CN1882083A (en) | Method for reducing image blocking effect |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |