CN102223534B - All-parallel bit plane coding method for image compression - Google Patents

Abstract

The invention relates to an all-parallel bit plane coding method for image compression, in particular to a high-speed bit plane coding method which is accorded with joint photographic experts group (JPEG) 2000 image compression standard, has low memory consumption and is in all-parallel coding of a bit plane and a coding channel. In the invention, single-row bits are used as a data unit, as long as significance status information of the bits at the previous row needs to be cached and primary data of the current row and the two following rows are read, coding operation of the channel at the current row and the bit plane can be carried out in one coding window; and then each time only a new column of data needs to be read, coding circulation can be realized; a realization structure of the coding method only needs 259 periods to process 32*32 wavelet sub bands.

Description

The full parallel bit plane coding method that is used for image compression

Technical field

The present invention relates to the full parallel bit plane coding method for image compression, particularly be applied to meet JPEG (joint photographic experts group) (JPEG) 2000 Standard of image compression, low memory consumption, the two-forty bit plane coding method of bit plane and the full parallel encoding of coding pass.

Background technology

At present international up-to-date image compression standard JPEG 2000 (list of references [1]) possess intraframe coding, compression effectiveness good, support the qualities such as encoding region of interest and progressive transmission, thereby in high-end digital picture is used, obtained popularization.In image compression process, at first adopt lift structure two-dimensional discrete wavelet conversion (DWT) original image, adopt afterwards bit plane encoder (BPC) to process the wavelet coefficient subband and generate context and code value (CX/D), send into again arithmetic encoder (AC) and carry out compressed encoding.Because BPC need to set up context to each bit of the wavelet transform result of image, computation complexity is high, and needs to consume great internal memory.Along with the application popularization of all kinds of high-definition digital images, how to improve the full degree of concurrence of bit plane coding method, and effectively reduce memory consumption and become one of this area research focus.

The BPC method that the JPEG2000 Standard of image compression adopts is proposed by David Taubman, P position after the method will quantize has the symbol wavelet coefficient to be organized as the code block of N * N size, then downward from sign bit, the symbol segmentation of wavelet coefficient is become an independently sign plane, the bit of wavelet coefficient sequentially is divided into the bit plane of (P-1) layer N * N size; In single bit plane, per 4 row of bit are divided into a band.Encoder is according to raster order, and from the highest order plane, from top to bottom from left to right, bit-plane by bit-plane is one by one with the coding by bit.

Wall scroll band Nepit position encryption algorithm is as follows:

1) judge whether current bit belongs to conspicuousness propagation ducts (SPP): if bit is current not remarkable, have at least one to be significant in 8 bits in territory and face, then it belongs to this passage, carries out Zero-code (ZC) and computational context.And then judge whether itself is remarkable; If then upgrade the conspicuousness state, carry out symbolic coding (SC); Otherwise do not carry out.If bit face the territory neither significantly or bit itself remarkable, then it does not belong to this passage, skips this point and removes to judge next point.All bits in the band are carried out this operation, then get back to starting point and carry out 2);

2) judge whether current bit belongs to amplitude refinement (MRP): if bit is significantly current and coding not, then belong to this passage, carry out amplitude refinement coding; Otherwise do not belong to this passage, skip this point and remove to judge next point.All bits in the band are carried out this operation, then get back to starting point and carry out 3);

3) carry out removing passage (CLP): use Run-Length Coding (RLC), Zero-code and symbolic coding to be encoded in all remaining bits positions that do not belong to conspicuousness propagation ducts and amplitude refinement.

Standard BP C algorithm is from the highest order plane, and from top to bottom from left to right, bit-plane by bit-plane is with one by one by the bit coding, and pursues channel coding according to the order of SPP passage, MRP passage, CLP passage.This sequential encoding mode speed is very slow, and storage needs a large amount of intermediate data, and the therefore coding rate of Effective Raise bit plane encoder and reduce storage inside consumption how becomes one of academic focus that Chinese scholars pays close attention to.

The representative achievement in research in this field has: document [1] has proposed the Bit-Plane Encoding implementation algorithm based on pixel the earliest, and is adopted by the JPEG2000 standard; Document [2] adopts the state machine structure to realize bit-plane coding; Document [3] proposes per-column coding pass parallel organization, namely adopts two window code devices that 3 passages are carried out scanning encoding simultaneously.Have sequencing because passage scans, after first window conspicuousness state renewal, second window just can begin coding by the time, and needing between two window code devices has increased the control complexity alternately; Document [4] has proposed that bit is skipped and the row skip algorithm, namely skips bit and the row of no context output, to reduce calculation times.The method can't be brought into play the advantage of circuit parallel, and degree of optimization depends on view data, the performance big rise and fall.Document [5] has proposed the bit plane parallel algorithm, and is parallel between the bit plane of having realized encoding.But the coding unit of single bit plane has adopted the scheme of document [3] and [4] simultaneously, and the control complexity is higher, and consumption of natural resource is larger, and this shortcoming is particularly outstanding after carrying out hardware multiplexing realizing the bit plane parallel encoding.

[1]Taubman?D.High?performance?scalable?image?compression?with?EBCOT.IEEE?Trans.on?Image?Processing，2000，9(7)：1158-1170

[2]Andra?K.，Chakrabarti?C.，Acharya?T.，A?high-performance?JPEG?2000?architecture.IEEE?Trans.on?Circuits?and?System?for?Video?Technology，2003，13(3)：209-218

[3]Chiang?J.-S.，Lin?Y.-S.，Hsieh?C.-Y.，Efficient?pass-parallel?architecture?for?EBCOT?in?JPEG?2000.In：IEEE?International?Symposium?on?Circuits?and?Systems，2002.773-776

[4]Lian?C.-J.，Chen?K.-F.，Chen?H.-H.，et?al.Analysis?and?architecture?design?of?block-coding?engine?for?EBCOT?in?JPEG?2000.IEEE?Trans.on?Circuits?and?System?for?Video?Technology，2003，13(3)：219-230

[5]Liu?K.，Li?Y.-S.，Wu?C.-K.，A?high?performance?EBCOT?coding?and?its?VLSI?architecture.Journal?of?Software，2006，17(7)：1553-1560

Summary of the invention

Technical problem: a kind of full parallel bit plane coding method for image compression that the present invention relates to, by removing the correlation between bit plane, realize the parallel absolute coding of each bit plane; In single bit plane, as data cell, the conspicuousness state information of buffer memory previous column bit only, and read initial data when prostatitis and follow-up two row is finished passage and bit plane parallel encoding when the prostatitis in a coding window with single-row bit; Only need read in afterwards the new data of row at every turn, realize the coding circulation.

The described correlation of having removed between bit plane, realize the parallel absolute coding of each bit plane, after referring to read in a row wavelet coefficient, change based on formula 1 and 2 pairs of wavelet coefficients that read in, generate conspicuousness state s and the first amplitude refinement state mrf on outer all the bit planes of divided-by symbol bit plane, thereby removed the correlation between bit plane; Each bit plane uses s and the mrf that generates, and gets final product the coding of independent parallel;

$s_l^n=\underset{i=l+1}{\overset{\mathrm{MSB}}{\mathrm{\Σ}}}{\mathrm{mag}}_i^n$ Formula 1

In formula 1 and the formula 2,

The conspicuousness state that represents the n row bit of l bit plane,

The first amplitude refinement state that represents the n row bit of l bit plane,

Represent the amplitude of i bit plane n row bit, MSB represents the number of plies on highest order plane, and ∑ represents exclusive disjunction.

Described in single bit plane, with single-row bit as data cell, the conspicuousness state information of buffer memory previous column bit only, and read initial data when prostatitis and follow-up two row, in a coding window, finish passage and bit plane parallel encoding when the prostatitis; Only need read in afterwards the new data of row at every turn, realize the coding circulation, refer to:

Standard bit plane coding BPC algorithm dictates, when n row bit in the bit-plane was carried out the SPP judgement, the conspicuousness state of itself and front and back adjacent column was respectively

Wherein

The conspicuousness state of expression (n-1) row bit after conspicuousness propagation ducts SPP coding upgrades,

Represent the original conspicuousness state of n row bit; When n row bit was carried out amplitude refinement MRP judgement, itself and front and back row conspicuousness state were respectively

When adjudicating remove channel C LP when the prostatitis, itself and front and back row conspicuousness state are respectively

If bit belongs to SPP or CLP, then may adopt Zero-code and symbolic coding to encode, thus the conspicuousness state of update bit position; If bit belongs to MRP, then must adopt the amplitude refinement to encode, can the conspicuousness state of bit not upgraded; Push away to get formula 3:

$s_{\mathrm{update}\_\mathrm{mrp}}^n=s_{\mathrm{update}\_\mathrm{spp}}^n$ Formula 3

Can be summarized as according to formula 3, the n row bits conspicuousness state update condition before and after three passage judgements:

$\left\{\begin{array}{c}{s}_{\mathrm{update}\_\mathrm{spp}}^n=f(s_{\mathrm{update}\_\mathrm{spp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1})\\ s_{\mathrm{update}\_\mathrm{mrp}}^n=f(s_{\mathrm{update}\_\mathrm{spp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1},s_{\mathrm{original}}^{n+2})\\ s_{\mathrm{update}\_\mathrm{clp}}^n=f(s_{\mathrm{update}\_\mathrm{clp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1},s_{\mathrm{original}}^{n+2})\end{array}\right.n\≥1$ Formula 4

In the formula 4,

Be used for the conspicuousness state after prediction (n+1) is listed in the SPP coding

Standard bit plane encryption algorithm regulation, the conspicuousness state beyond the border, the band left and right sides is defaulted as 0, and then the conspicuousness state update condition of the 0th row bit before and after three channel codings is:

$\left\{\begin{array}{c}{s}_{\mathrm{update}\_\mathrm{<figure-callout\; id="0"\; label="spp"\; filenames="110818112806.png,110818112809.png"\; state="\{\{state\}\}">spp</figure-callout>}}^0=f(0,s_{\mathrm{original}}^0,s_{\mathrm{original}}^1)\\ s_{\mathrm{update}\_\mathrm{<figure-callout\; id="0"\; label="mrp"\; filenames="110818112806.png,110818112809.png"\; state="\{\{state\}\}">mrp</figure-callout>}}^0=f(0,s_{\mathrm{original}}^0,s_{\mathrm{original}}^1,s_{\mathrm{original}}^2)\\ s_{\mathrm{update}\_\mathrm{<figure-callout\; id="0"\; label="clp"\; filenames="110818112806.png,110818112809.png"\; state="\{\{state\}\}">clp</figure-callout>}}^0=f(0,s_{\mathrm{original}}^0,s_{\mathrm{original}}^1,s_{\mathrm{original}}^2)\end{array}\right.n=0$ Formula 5

Reading n row, (n+1) row behind the wavelet coefficient of (n+2) row, obtain each bit planes by formula 1 and formula 2

Consist of a coding window; If known (n-1) is listed as simultaneously

With

As supplementary column, can finish the coding pass parallel encoding of n row according to formula 3 and formula 4, and will

With

Buffer memory is as new supplementary column; Afterwards, coding window only need read in the circulation that the new wavelet coefficient of row can be realized encryption algorithm; First required during coding window initial code conspicuousness state supplementary then can be tried to achieve by formula 5.

Beneficial effect: this algorithm has been realized bit plane and channel parallel simultaneously, and its superiority is to be easy to the corresponding hardware pipeline operation that realizes; Only need a coding window, calculate and control simply, memory space is little, can keep less resource overhead when hardware is realized; Can effectively reduce the input-output operation complexity between DWT and the BPC.After BPC takes out three row wavelet coefficients, can finish simultaneously first and calculate at CX/D of all bit planes, and will work as behind the SPP coding in prostatitis and the conspicuousness state cache of CLP after encoding.Afterwards, the new wavelet coefficient of row is read in each circulation, can carry out the Bit-Plane Encoding of next column.Corresponding pile line operation very easily realizes when hardware is realized.In single bit plane, only need store the required auxiliary parameter of present encoding window, need not more buffer memory consumption.Coding window is only encoded to a column data at every turn, and amount of calculation is little, and data read dependence pile line operation, and the control complexity is low.Therefore, when realizing, hardware can keep than low consumption of resources.Because when realizing the bit plane parallel encoding, need to carry out the hardware cell of single bit plane encoder multiplexingly, so this quality is particularly important.Adopt this kind algorithm after reading in three row wavelet coefficients, but BPC finish coding and output from the just immediate operation of every taking-up one column data of DWT, need not unnecessary storage and control, has greatly improved the efficient of the data input and output between BPC and the DWT.

Description of drawings

The present invention will be further described below in conjunction with drawings and Examples.

In Figure 1B PC algorithm, band, bit are according to raster order scanning encoding schematic diagram.

Fig. 2 realizes three coding pass parallel encoding passage principle schematic by a coding window.

Fig. 3 bit plane parallel encoding schematic diagram.

The full parallel bit plane coding method implementation structure of Fig. 4 schematic diagram.

Embodiment

The embodiment that provides below in conjunction with accompanying drawing and inventor is described in further details the present invention.

The present invention has at first removed the correlation between each bit plane, namely obtains each bit plane conspicuousness state s and first amplitude refinement state mrf by the state parameter generation unit; After the register buffer unit buffers, in single bit plane, the conspicuousness state information of buffer memory previous column bit, and read when prostatitis and follow-up two initial data that are listed as, passage and bit plane parallel encoding when the prostatitis just can in a coding window, be finished; Only need read in afterwards the new data of row at every turn, can realize the coding circulation.

The implementation example is as follows:

1) reads in a row wavelet coefficient, change based on formula 1 and 2 pairs of wavelet coefficients that read in, the conspicuousness state of all bit planes of generation except sign bit plane and highest order plane and first amplitude refinement state send into that each bit plane is parallel encodes; The coding structure of each bit plane is the same with cataloged procedure;

2) (n-1) row conspicuousness state parameter, first amplitude refinement state parameter, amplitude and the symbol finished of coding window output buffer memory is as auxiliary, and read in new (n+2) conspicuousness state parameter, first amplitude refinement state parameter, amplitude and symbol, add existing n row and (n+1) row in the coding window, begin the n row are encoded;

3) judge passage under each bit according to the conspicuousness state parameter, if certain bit conspicuousness is " 0 ", appoint in the bit that is adjacent have a bit conspicuousness for " 1 " then this bit belong to SPP, if certain bit conspicuousness is " 1 ", then this bit belongs to MRP, other bits belong to CLP, after judgement is finished each bit are encoded according to different passages, and the coding of each passage walks abreast;

4) for the SPP passage, be that the bit of " 0 " uses ZC to calculate CX/D to amplitude, be that the bit of " 1 " uses ZC and SC to calculate CX/D to amplitude, upgrade the conspicuousness of each bit that has carried out the SPP passage and deposit, when the next column data are carried out the SPP passage; For the MRP passage, predict that at first each bit is through the conspicuousness after the SPP passage, and the conspicuousness of each bit upgraded, then use MRC to calculate the CX/D of each bit, the conspicuousness of each bit of MRP passage has been carried out in renewal, when the next column data are carried out the MRP passage; For the CLP passage, predict that at first each bit is through the conspicuousness after the SPP passage, and the conspicuousness of each bit upgraded, then be " 0 " for conspicuousness own, and when the bit conspicuousness that is adjacent also is " 0 " entirely, use RLC to calculate CX/D, for other bits, if this bit amplitude is " 0 ", use ZC to calculate CX/D, if amplitude is " 1 ", use ZC and SC to calculate CX/D, the conspicuousness of each bit of CLP passage has been carried out in renewal, when the next column data are carried out the CLP passage.

5) to after the end-of-encode of n row, return 1), begin to read in new row and continue loop coding.

Claims (2)

1. a full parallel bit plane coding method that is used for image compression is characterized in that removing the correlation of using the bit plane that obtains in the JPEG2000 image compression process, realizes the parallel absolute coding of each bit plane; In single bit plane, as data cell, the conspicuousness state information of buffer memory previous column bit only, and read initial data when prostatitis and follow-up two row is finished passage and bit plane parallel encoding when the prostatitis in a coding window with single-row bit; Only need read in afterwards the new data of row at every turn, realize the coding circulation;

The described correlation of having removed between bit plane, realize the parallel absolute coding of each bit plane, after referring to read in a row wavelet coefficient, change based on formula 1 and 2 pairs of wavelet coefficients that read in, generate conspicuousness state s and the first amplitude refinement state mrf on divided-by symbol bit plane and out-of-plane all the bit planes of highest order, thereby removed the correlation between bit plane; Each bit plane uses s and the mrf that generates, and gets final product the coding of independent parallel;

$s_l^n=\underset{i=l+1}{\overset{\mathrm{MSB}}{\mathrm{\&Sigma;}}}{\mathrm{mag}}_i^n$ Formula 1

Formula 2

In formula 1 and the formula 2,

The conspicuousness state that represents the n row bit of l bit plane,

The first amplitude refinement state that represents the n row bit of l bit plane,

Represent the amplitude of i bit plane n row bit, MSB represents the number of plies on highest order plane, and Σ represents exclusive disjunction.

2. a kind of full parallel bit plane coding method for image compression as claimed in claim 1, it is characterized in that described in single bit plane, with single-row bit as data cell, the conspicuousness state information of buffer memory previous column bit only, and read when prostatitis and follow-up two initial data that are listed as, in a coding window, finish passage and bit plane parallel encoding when the prostatitis; Only need read in afterwards the new data of row at every turn, realize the coding circulation, refer to:

Standard bit plane coding BPC algorithm dictates, when n row bit in the bit-plane was carried out the SPP judgement, the conspicuousness state of itself and front and back adjacent column was respectively

, ,

Wherein

The conspicuousness state of expression (n-1) row bit after conspicuousness propagation ducts SPP coding upgrades,

Represent the original conspicuousness state of n row bit; When n row bit was carried out amplitude refinement MRP judgement, itself and front and back row conspicuousness state were respectively

,

,

When adjudicating remove channel C LP when the prostatitis, itself and front and back row conspicuousness state are respectively

,

,

If bit belongs to SPP or CLP, then may adopt Zero-code and symbolic coding to encode, thus the conspicuousness state of update bit position; If bit belongs to MRP, then must adopt the amplitude refinement to encode, can the conspicuousness state of bit not upgraded; Push away to get formula 3:

$s_{\mathrm{update}\_\mathrm{mrp}}^n=s_{\mathrm{update}\_\mathrm{spp}}^n$ Formula 3

Can be summarized as according to formula 3, the n row bits conspicuousness state update condition before and after three passage judgements:

$\left\{\begin{array}{c}{s}_{\mathrm{update}\_\mathrm{spp}}^n=f(s_{\mathrm{update}\_\mathrm{spp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1})\\ s_{\mathrm{update}\_\mathrm{mrp}}^n=f(s_{\mathrm{update}\_\mathrm{spp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1},s_{\mathrm{original}}^{n+2})\\ s_{\mathrm{update}\_\mathrm{clp}}^n=f(s_{\mathrm{update}\_\mathrm{clp}}^{n-1},s_{\mathrm{original}}^n,s_{\mathrm{original}}^{n+1},s_{\mathrm{original}}^{n+2})\end{array}\right.n\&GreaterEqual;1$ Formula 4

In the formula 4,

Be used for the conspicuousness state after prediction (n+1) is listed in the SPP coding

Standard bit plane encryption algorithm regulation, the conspicuousness state beyond the border, the band left and right sides is defaulted as 0, and then the conspicuousness state update condition of the 0th row bit before and after three channel codings is:

$\left\{\begin{array}{c}{s}_{\mathrm{update}\_\mathrm{spp}}^0=f(0,s_{\mathrm{original}}^n,s_{\mathrm{original}}^1)\\ s_{\mathrm{update}\_\mathrm{mrp}}^0=f(0,s_{\mathrm{original}}^0,s_{\mathrm{original}}^1,s_{\mathrm{original}}^2)\\ s_{\mathrm{update}\_\mathrm{clp}}^0=f(0,s_{\mathrm{original}}^0,s_{\mathrm{original}}^1,s_{\mathrm{original}}^2)\end{array}\right.n=0$ Formula 5

Reading n row, (n+1) row behind the wavelet coefficient of (n+2) row, obtain each bit planes by formula 1 and formula 2 ,

,

, consist of a coding window; If known (n-1) is listed as simultaneously

With As supplementary column, can finish the coding pass parallel encoding of n row according to formula 3 and formula 4, and will

With

Buffer memory is as new supplementary column; Afterwards, coding window only need read in the circulation that the new wavelet coefficient of row can be realized encryption algorithm; First required during coding window initial code conspicuousness state supplementary then can be tried to achieve by formula 5.

Images