CN106559668A - A kind of low code rate image compression method based on intelligent quantization technology - Google Patents
A kind of low code rate image compression method based on intelligent quantization technology Download PDFInfo
- Publication number
- CN106559668A CN106559668A CN201510621295.9A CN201510621295A CN106559668A CN 106559668 A CN106559668 A CN 106559668A CN 201510621295 A CN201510621295 A CN 201510621295A CN 106559668 A CN106559668 A CN 106559668A
- Authority
- CN
- China
- Prior art keywords
- prime
- image
- beta
- image block
- coding
- 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.)
- Granted
Links
Landscapes
- Compression Of Band Width Or Redundancy In Fax (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention provides a kind of low code rate image compression method based on intelligent quantization technology, it is on the premise of low-bit-rate compact, for internal smoother image block, using being encoded based on the method for intelligent quantization technology, rebuild in conjunction with interpolation technique, on the premise of control coding distortion, encoder bit rate is effectively reduced;Simultaneously for the more complicated image block of inner vein, ensure the high efficient coding to these image blocks using traditional coding method based on jpeg image compression standard.Different coding strategies are adaptive selected to the image block with different texture feature, the high efficient coding to whole image is realized.Compared with traditional JPEG image compression method, the present invention can overcome the single shortcoming of coding mode in conventional JPEG image compression method.
Description
Technical field
The invention belongs to image code domain, relates generally to the compress technique of digital picture.
Background technology
Low bit- rate compression of images is commonly applied in the scenes such as video monitoring, visual telephone, satellite remote sensing images transmission.It is adjoint
The extensive application of wireless communication networks, requirement more and more higher of the people to realtime graphic transmission under finite bandwidth, and it is efficient low
Code rate image compression technology, can meet carries out the demand of real-time Transmission to image under low-bandwidth condition.
Method for compressing image can be divided into two class of Lossless Compression and lossy compression method, conventional compression algorithm be by conversion,
Several core technologies such as quantization and entropy code, realize the lossy compression method to picture signal.And in order to realize efficient low bit- rate figure
As compression, traditional method is, under existing Image Lossy Compression framework, to reduce the compression quality of image as cost, to reach
Relatively low encoder bit rate.
For natural image, typically it is made up of complex texture regional peace skating area domain inside image, for texture ratio
More complicated region, using traditional coding strategy, when no matter for high Compression or low-bit-rate compact, all more
Effectively.And for internal smoother region, high-quality compression can be carried out by selected section pixel in compression process, and
Low quality compression is carried out to rest of pixels point, with the code check for controlling to encode.After the completion of compressed encoding, the picture compressed with high-quality
Vegetarian refreshments can realize the reconstruction of whole smooth region by the method for interpolation.Such coding strategy not only can effective control volume
Code code check, and will not cause larger coding distortion in smooth region, particularly higher in compression ratio, i.e., encoder bit rate compared with
In the case of low, to controlling encoder bit rate and coding distortion degree, can be much more efficient.Therefore, in the case where encoder bit rate is relatively low
Different coding strategies is adopted to the zones of different inside image, it will code efficiency is greatly improved.Traditional compression of images side
Method is encoded using identical strategy to image inside zones of different, it is impossible to improve local code according to the provincial characteristics of image
Efficiency, thus results in binary encoding inefficiency, referring to bibliography " JPEG (Joint Photographic Experts
Group):ISO/IEC IS 10918–1/ITU-T Recommendation T.81,Digital Compression and
Coding of Continuous-Tone Still Image,1993”。
The content of the invention
It is an object of the invention to provide a kind of low code rate image compression method based on intelligent quantization technology, it is in low code
Under the coding environment of rate, each image block is allowed to be adaptive selected a kind of coding mode by rate distortion optimized strategy, so as to have
The code efficiency of image block is pointedly improved, the Efficient Compression to whole image signal under the conditions of low bit- rate is realized.With tradition
JPEG image compression method compare, the invention provides two kinds of coding modes under the conditions of low bit- rate, can overcome tradition
The single shortcoming of coding mode in JPEG image compression method.
Present disclosure is described for convenience, does following term definition first:
1 is defined, the method for image block in traditional jpeg image compression standard
Method of traditional image block method according to piecemeal is carried out to image in Joint Photographic Experts Group, original image is divided into
The equidimension image block of multiple non-overlapping copies, specifically describes process referring to " JPEG (Joint Photographic Experts
Group):ISO/IEC IS 10918–1/ITU-T Recommendation T.81,Digital Compression and
Coding of Continuous-Tone Still Image,1993”;
2 are defined, traditional method for compressing image based on intelligent quantization technology
Traditional method for compressing image based on intelligent quantization technology can as needed to the part in each image block
Pixel carries out high-quality compression, carries out low-quality compression to remaining pixel.The method will carry out high-quality compression
The coordinated indexing set of pixel be defined as Ω1, the coordinated indexing set for carrying out the pixel of low quality compression is defined as
Ω2;And conversion coefficient is divided into into two groups in transform domain is commonly quantified respectively and pressure type quantifies, common quantization will be carried out
The coordinated indexing set of conversion coefficient is defined as Ψ1, the coordinated indexing set for carrying out the conversion coefficient of pressure type quantization is defined
For Ψ2;Meanwhile, in traditional intelligent quantization technology provide three kinds of quantization strategies, i.e. quantization strategy -1, quantization strategy -2 and
Quantization strategy -3, as selection when quantifying.For the image block of input, the method can be encoded and be decoded, Yi Jiji
Calculate the number of coded bits of encoded images;Specific descriptions process is referring to document " Constrained quantization in the
transform domain with applications in arbitrarily-shaped object coding”;
3 are defined, traditional method for compressing image based on JPEG coding standards
Traditional method for compressing image based on JPEG coding standards can realize the coding to image and decoding, Yi Jiji
Calculate the number of coded bits of encoded images;Specific descriptions process is referring to " JPEG (Joint Photographic Experts
Group):ISO/IEC IS 10918–1/ITU-T Recommendation T.81,Digital Compression and
Coding of Continuous-Tone Still Image,1993”;
4 are defined, traditional bicubic interpolation method
Traditional bicubic interpolation method is the most frequently used interpolation method in two-dimensional space, in this interpolation method, point
The value at (u, v) place can be obtained by the weighted average of 16 points nearest in rectangular mesh around it;Specific descriptions process
Referring to document " Cubic convolution interpolation for digital image processing ";
5 are defined, traditional calculating mean square error methodology
Traditional calculating mean square error methodology for two sizes be m × n two dimensional input signal X andAccording to the following formula
Calculate the mean square error between them:
6 are defined, the method for image block composograph in traditional jpeg image compression standard
The method of traditional image block composograph is according to being carried out with image block in jpeg image compression standard mutually not
Method of the overlapping combinations to synthesize complete image, specifically describes process referring to " JPEG (Joint Photographic Experts
Group):ISO/IEC IS 10918–1/ITU-T Recommendation T.81,Digital Compression and
Coding of Continuous-Tone Still Image,1993”;
The invention provides a kind of low code rate image compression method based on intelligent quantization technology, it comprises the following steps:
Step 1, the pretreatment of image
Be the image of W × H by size, N is divided into according to the method for image block in traditional jpeg image compression standard
=(W × H)/82Individual non-overlapping copies, size is 8 × 8 square image blocks, is designated as B1, B2..., Bi..., BN, here, W generations
The width of table image, the height of H representative images, the total number of image block after the division of N representative images, the index of i representative image blocks,
I ∈ { 1,2 ..., N };
Step 2, the generation of index matrix
64 natural numbers 1,2 ..., 64 by from small to large, and order from top to bottom is put by column, produces a size and is
8 × 8 index matrix, is designated as I:
Element in I is designated as I (x, y), and (x and y are natural numbers, and 1≤x≤8,1≤y≤8), and here, x represents rope
Draw the abscissa of matrix I interior elements, y represents the ordinate of index matrix I interior elements;
Step 3, the image block coding parameter based on intelligent quantization technology are arranged
First, using the quantization strategy -2 provided in traditional encryption algorithm based on intelligent quantization technology;
Secondly, need to carry out in defining traditional encryption algorithm based on intelligent quantization technology the pixel of high-quality compression
Coordinated indexing collection is combined intoNeed to carry out the pixel of low quality compression in defining traditional encryption algorithm based on intelligent quantization technology
The coordinated indexing collection of point is combined intoHere,It is the row vector of 1 × 32,For even number, and 1≤
X≤8,1≤y≤8 }, i.e., It is the row vector of 1 × 32,For odd number, and 1≤x≤8,1≤y≤8 }, i.e., Wherein, I is the index matrix produced in step 2, x generations
The abscissa of table index matrix I interior elements, y represent the ordinate of index matrix I interior elements, and x and y is natural number;
Finally, need to carry out the common conversion coefficient for quantifying in defining traditional encryption algorithm based on intelligent quantization technology
Coordinated indexing collection is combined intoNeed to carry out the change of pressure type quantization in defining traditional encryption algorithm based on intelligent quantization technology
The coordinated indexing collection for changing coefficient is combined intoHere,It is the row vector of 1 × 32, It is the row vector of 1 × 32,
Step 4, the image block based on intelligent quantization technology are encoded
First, by the size produced by step 1 for 8 × 8 image block Bi, according to the tradition being provided with step 3
Encoded and decoded based on the method for compressing image of intelligent quantization technology, the bit number after being encoded is designated asWith
And decoded image block, it is designated as B'i,
I.e.
Here, β 'm,nIt is B'iIn element, m represents B'iThe abscissa of interior element, n represent B'iThe ordinate of interior element, m
It is natural number with n, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N represent image in step 1
The total number of image block after division;
Then, with traditional bicubic interpolation method to B'iIn be located at (u, v) position on pixel enter row interpolation, obtain
Reconstruction image block to after interpolation, is designated as
Here, u is B'iThe abscissa of interior pixel, v are B'iThe ordinate of interior pixel, u and v are natural numbers, 1≤u≤
8,1≤v≤8, and u+v is odd number;It isIn element, m representThe abscissa of interior element, n are representedInterior element
Ordinate, m and n is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N ride instead of walk
The total number of image block after image is divided in rapid 1;
Finally, image block is calculated with traditional calculations mean square error methodologyWith the image block B produced in step 1iBetween
Mean square error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Step 5, encodes original picture block with traditional coding method
First, by the size produced by step 1 for 8 × 8 image block Bi, according to traditional based on JPEG coding standards
Method for compressing image encoded and decoded, the bit number after being encoded is designated asAnd decoded image block,
It is designated as B "i,
I.e.
Here, β "m,nIt is B "iIn element, m represents B "iThe abscissa of interior element, n represent B "iThe ordinate of interior element, m
It is natural number with n, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N represent image in step 1
The total number of image block after division;
Then, image block B is calculated with traditional calculations mean square error methodology "iWith the image block B produced in step 1iBetween
Mean square error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
The selection of step 6, optimum code and decoding schema
The image block B produced by step 1i, the bit number obtained with step 4And mean square errorIt is multiplied, will
The Coding cost obtained after multiplication is designated as The bit number obtained with step 5With it is equal
Square errorIt is multiplied, the Coding cost obtained after multiplication is designated as
RelativelyWithSize, ifThe coding and decoding methods of step 4 are selected so
To the image block B produced in step 1iEncoded and decoded, the code check after coding isIfThe coding and decoding methods of step 5 are selected so to the image block B that produces in step 1iCarry out coding reconciliation
Yard, the code check after coding is
Image block will be obtained after decoding, be designated as bi,
Here, αm,nIt is biIn element, m represents biThe abscissa of interior element, n represent biThe ordinate of interior element, m and n
It is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Step 7, reconstruction image
For the reconstruction image block b produced in step 6i, synthesized using image block in traditional jpeg image compression standard
The method of image, produces reconstruction image, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, N represents step
The total number of image block after image is divided in 1.
The general principle of the present invention:Intelligent quantization technology is that the high-quality compression for realizing image block interior section pixel is carried
Technical support is supplied.On the premise of low-bit-rate compact, for internal smoother image block, using based on intelligent quantization skill
The method of art is encoded, and is rebuild in conjunction with interpolation technique, on the premise of control coding distortion effectively can be reduced compiling
Code code check;Simultaneously for the more complicated image block of inner vein, traditional coding method based on jpeg image compression standard
Can ensure that the high efficient coding to these image blocks.Different volumes are adaptive selected to the image block with different texture feature
Code strategy, it is possible to achieve the high efficient coding to whole image.
The present invention essence be:Coding method based on intelligent quantization technology is mutually tied with traditional JPEG coding methods
Close, under conditions of low-bit-rate compact, different coding strategies adopted to different images block adaptive to realize optimum code,
So as to reach the Efficient Compression to whole image.
The innovative point of the present invention:The present invention is applied to intelligent quantization technology in low bit- rate compression of images a kind of new to define
The coding mode of type, by rate distortion optimized strategy, allows each image block to be adaptive selected coding mode, realizes optimum volume
Code, so as to reach the purpose for improving whole image compression efficiency.
Advantages of the present invention:Intelligent quantization technology is applied to into the coding in image smoothing region, it is relatively low in encoder bit rate
In the case of, with effective control coding distortion and encoder bit rate can be reduced.By this coding method with traditional based on jpeg image pressure
The coding method of contracting standard combines, and can improve the overall compression efficiency of image.
Description of the drawings
Fig. 1 be the present invention realize flow process;
Fig. 2 is the PSNR values obtained under identical encoder bit rate using different images coding method.
Specific embodiment
The feasibility that the system model is mainly verified by the way of emulation experiment of the invention, all steps all pass through experiment
Checking, is to realize the compression of images based on transform domain down-sampling technology, and specific implementation step is as follows:
Step 1, the pretreatment of image
The width W=8 of setting imagem, the height H=8 of imagen, m and n are natural numbers here, according to traditional JPEG
In Standard of image compression, the method for image block is divided into N=(W × H)/82Individual non-overlapping copies, size is 8 × 8 square
Image block, is designated as B1, B2..., Bi..., BN, here, the width of W representative images, the height of H representative images, N representative images are drawn
The total number of image block, the index of i representative image blocks, i ∈ { 1,2 ..., N } after point;
Step 2, the generation of index matrix
64 natural numbers 1,2 ..., 64 by from small to large, and order from top to bottom is put by column, produces a size and is
8 × 8 index matrix, is designated as I:
Element in I is designated as I (x, y), and (x and y are natural numbers, and 1≤x≤8,1≤y≤8), and here, x represents rope
Draw the abscissa of matrix I interior elements, y represents the ordinate of index matrix I interior elements;
Step 3, the image block coding parameter based on intelligent quantization technology are arranged
First, using the quantization strategy -2 provided in traditional encryption algorithm based on intelligent quantization technology;
Secondly, the seat of the pixel for needing to carry out high-quality compression in defining traditional encryption algorithm based on intelligent quantization technology
Mark indexed set is combined intoNeed to carry out the pixel of low quality compression in defining traditional encryption algorithm based on intelligent quantization technology
Coordinated indexing collection be combined intoHere,It is the row vector of 1 × 32,For even number, and 1
≤ x≤8,1≤y≤8 }, i.e., It is the row vector of 1 × 32,For odd number, and 1≤x≤8,1≤y≤8 }, i.e., Wherein, I is the index matrix produced in step 2, x generations
The abscissa of table index matrix I interior elements, y represent the ordinate of index matrix I interior elements, and x and y is natural number;
Finally, need to carry out the common conversion coefficient for quantifying in defining traditional encryption algorithm based on intelligent quantization technology
Coordinated indexing collection is combined intoNeed to carry out the change of pressure type quantization in defining traditional encryption algorithm based on intelligent quantization technology
The coordinated indexing collection for changing coefficient is combined intoHere,It is the row vector of 1 × 32, It is the row vector of 1 × 32,
Step 4, the image block based on intelligent quantization technology are encoded
First, by the size produced by step 1 for 8 × 8 image block Bi, according to the tradition being provided with step 3
Encoded and decoded based on the method for compressing image of intelligent quantization technology, the bit number after being encoded is designated as Biti (1),
And decoded image block, it is designated as B'i,
I.e.
Here, β 'm,nIt is B'iIn element, m represents B'iThe abscissa of interior element, n represent B'iThe ordinate of interior element, m
It is natural number with n, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N represent image in step 1
The total number of image block after division;
Then, with traditional bicubic interpolation method to B'iIn be located at (u, v) position on pixel enter row interpolation, obtain
Reconstruction image block to after interpolation, is designated as
Here, u is B'iThe abscissa of interior pixel, v are B'iThe ordinate of interior pixel, u and v are natural numbers, 1≤u≤
8,1≤v≤8, and u+v is odd number;It isIn element, m representThe abscissa of interior element, n are representedInterior unit
The ordinate of element, m and n is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N are represented
The total number of image block after image is divided in step 1;
Finally, image block is calculated with traditional calculations mean square error methodologyWith the image block B produced in step 1iBetween
Mean square error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Step 5, encodes original picture block with traditional coding method
First, by the size produced by step 1 for 8 × 8 image block Bi, according to traditional based on JPEG coding standards
Method for compressing image encoded and decoded, the bit number after being encoded is designated asAnd decoded image block,
It is designated as B "i,
I.e.
Here, β "m,nIt is B "iIn element, m represents B "iThe abscissa of interior element, n represent B "iThe ordinate of interior element, m
It is natural number with n, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N represent image in step 1
The total number of image block after division;
Then, image block B is calculated with traditional calculations mean square error methodology "iWith the image block B produced in step 1iBetween
Mean square error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
The selection of step 6, optimum code and decoding schema
The image block B produced by step 1i, the bit number obtained with step 4And mean square errorIt is multiplied, will
The Coding cost obtained after multiplication is designated as The bit number obtained with step 5With it is equal
Square errorIt is multiplied, the Coding cost obtained after multiplication is designated as RelativelyWithSize, ifThe coding and decoding methods of step 4 are selected so in step 1
The image block B of generationiEncoded and decoded, the code check after coding isIfSo select
The coding and decoding methods of step 5 are selected to the image block B that produces in step 1iEncoded and decoded, the code check after coding
ForImage block will be obtained after decoding, be designated as bi,
Here, αm,nIt is biIn element, m represents biThe abscissa of interior element, n represent biThe ordinate of interior element, m and n
It is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Step 7, reconstruction image
For the reconstruction image block b produced in step 6i, synthesized using image block in traditional jpeg image compression standard
The method of image, produces reconstruction image, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, N represents step
The total number of image block after image is divided in 1.
In the classical legend that embodiment is applied to Lena and two width resolution ratio of Barbara for 512 × 512, accompanying drawing 2 be
Under different encoder bit rates, the peak value letter that the method for compressing image different to different images application is encoded and obtained after being decoded
Make an uproar than (peak signal to noise ratio, PSNR).It is obvious that the method in the present invention has significantly than existing methods
Performance boost.
Claims (1)
1. a kind of low code rate image compression method based on intelligent quantization technology, is characterized in that it comprises the following steps:
Step 1, the pretreatment of image
Be the image of W × H by size, N=(W are divided into according to the method for image block in traditional jpeg image compression standard
×H)/82Individual non-overlapping copies, size is 8 × 8 square image blocks, is designated as B1, B2..., Bi..., BN, here, W representative graphs
The width of picture, the height of H representative images, the total number of image block, the index of i representative image blocks, i ∈ after the division of N representative images
{ 1,2 ..., N };
Step 2, the generation of index matrix
64 natural numbers 1,2 ..., 64 by from small to large, and order from top to bottom is put by column, and it is 8 × 8 to produce a size
Index matrix, be designated as I:
Element in I is designated as I (x, y), and (x and y are natural numbers, and 1≤x≤8,1≤y≤8), and here, x represents index square
The abscissa of battle array I interior elements, y represent the ordinate of index matrix I interior elements;
Step 3, the image block coding parameter based on intelligent quantization technology are arranged
First, using the quantization strategy -2 provided in traditional encryption algorithm based on intelligent quantization technology;
Secondly, the coordinate of the pixel for needing to carry out high-quality compression in defining traditional encryption algorithm based on intelligent quantization technology
Indexed set is combined intoNeed to carry out the pixel of low quality compression in defining traditional encryption algorithm based on intelligent quantization technology
The coordinated indexing collection of point is combined intoHere,It is the row vector of 1 × 32, I.e. It is the row vector of 1 × 32,I.e. Wherein, I is the index matrix produced in step 2, x generations
The abscissa of table index matrix I interior elements, y represent the ordinate of index matrix I interior elements, and x and y is natural number;
Finally, need to carry out the coordinate of the common conversion coefficient for quantifying in defining traditional encryption algorithm based on intelligent quantization technology
Indexed set is combined intoNeed to carry out the transformation series of pressure type quantization in defining traditional encryption algorithm based on intelligent quantization technology
Several coordinated indexing collection are combined intoHere,It is the row vector of 1 × 32, It is the row vector of 1 × 32,
Step 4, the image block based on intelligent quantization technology are encoded
First, by the size produced by step 1 for 8 × 8 image block Bi, according to be provided with step 3 it is traditional based on
The method for compressing image of intelligent quantization technology is encoded and is decoded, the bit number after being encoded, and is designated asAnd decoding
Image block afterwards, is designated as B'i,
I.e.
Here, β 'm,nIt is B'iIn element, m represents B'iThe abscissa of interior element, n represent B'iThe ordinate of interior element, m and n
It is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Then, with traditional bicubic interpolation method to B'iIn be located at (u, v) position on pixel enter row interpolation, obtain interpolation
Reconstruction image block afterwards, is designated as
Here, u is B'iThe abscissa of interior pixel, v are B'iThe ordinate of interior pixel, u and v are natural numbers, 1≤u≤8,1
≤ v≤8, and u+v is odd number;It isIn element, m representThe abscissa of interior element, n are representedInterior element
Ordinate, m and n are natural numbers, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N represents step
The total number of image block after image is divided in 1;
Finally, image block is calculated with traditional calculations mean square error methodologyWith the image block B produced in step 1iBetween it is square
Error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, N are represented after image is divided in step 1 and are schemed
As the total number of block;
Step 5, encodes original picture block with traditional coding method
First, by the size produced by step 1 for 8 × 8 image block Bi, according to traditional figure based on JPEG coding standards
As compression method is encoded and is decoded, the bit number after being encoded is designated asAnd decoded image block, it is designated as
B”i,
I.e.
Here, β "m,nIt is B "iIn element, m represents B "iThe abscissa of interior element, n represent B "iThe ordinate of interior element, m and n
It is natural number, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, during N represents step 1, image is divided
The total number of image block afterwards;
Then, image block B is calculated with traditional calculations mean square error methodology "iWith the image block B produced in step 1iBetween it is square
Error, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, N are represented after image is divided in step 1 and are schemed
As the total number of block;
The selection of step 6, optimum code and decoding schema
The image block B produced by step 1i, the bit number obtained with step 4And mean square errorIt is multiplied, will be multiplied
The Coding cost for obtaining afterwards is designated asThe bit number obtained with step 5It is square
ErrorIt is multiplied, the Coding cost obtained after multiplication is designated as
RelativelyWithSize, ifThe coding and decoding methods of step 4 are selected so to step
The image block B produced in rapid 1iEncoded and decoded, the code check after coding isIfThat
The coding and decoding methods of step 5 are selected to the image block B that produces in step 1iEncoded and decoded, the code after coding
Rate is
Image block will be obtained after decoding, be designated as bi,
Here, αm,nIt is biIn element, m represents biThe abscissa of interior element, n represent biThe ordinate of interior element, m and n are certainly
So count, 1≤m≤8,1≤n≤8;The index of i representative image blocks, i ∈ { 1,2 ..., N }, N are represented after image is divided in step 1 and are schemed
As the total number of block;
Step 7, reconstruction image
For the reconstruction image block b produced in step 6i, using image block composograph in traditional jpeg image compression standard
Method, produces reconstruction image, is designated asHere, the index of i representative images block, i ∈ { 1,2 ..., N }, N are schemed in representing step 1
The total number of image block after as dividing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510621295.9A CN106559668B (en) | 2015-09-25 | 2015-09-25 | A kind of low code rate image compression method based on intelligent quantization technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510621295.9A CN106559668B (en) | 2015-09-25 | 2015-09-25 | A kind of low code rate image compression method based on intelligent quantization technology |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106559668A true CN106559668A (en) | 2017-04-05 |
CN106559668B CN106559668B (en) | 2018-07-27 |
Family
ID=58414399
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510621295.9A Active CN106559668B (en) | 2015-09-25 | 2015-09-25 | A kind of low code rate image compression method based on intelligent quantization technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106559668B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108737976A (en) * | 2018-05-22 | 2018-11-02 | 南京大学 | A kind of compression transmitting method based on Big Dipper short message |
CN110113609A (en) * | 2019-04-26 | 2019-08-09 | 深圳市华星光电技术有限公司 | Method for compressing image and device |
CN110717875A (en) * | 2019-10-18 | 2020-01-21 | 华北理工大学 | High-definition image processing method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101299819A (en) * | 2008-04-25 | 2008-11-05 | 清华大学 | Method for sorting three-dimensional wavelet sub-band and enveloping code flow of telescopic video coding |
CN101408769A (en) * | 2008-11-21 | 2009-04-15 | 冶金自动化研究设计院 | On-line energy forecasting system and method based on product ARIMA model |
US20120170668A1 (en) * | 2011-01-04 | 2012-07-05 | The Chinese University Of Hong Kong | High performance loop filters in video compression |
-
2015
- 2015-09-25 CN CN201510621295.9A patent/CN106559668B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101299819A (en) * | 2008-04-25 | 2008-11-05 | 清华大学 | Method for sorting three-dimensional wavelet sub-band and enveloping code flow of telescopic video coding |
CN101408769A (en) * | 2008-11-21 | 2009-04-15 | 冶金自动化研究设计院 | On-line energy forecasting system and method based on product ARIMA model |
US20120170668A1 (en) * | 2011-01-04 | 2012-07-05 | The Chinese University Of Hong Kong | High performance loop filters in video compression |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108737976A (en) * | 2018-05-22 | 2018-11-02 | 南京大学 | A kind of compression transmitting method based on Big Dipper short message |
CN108737976B (en) * | 2018-05-22 | 2021-05-04 | 南京大学 | Compression transmission method based on Beidou short message |
CN110113609A (en) * | 2019-04-26 | 2019-08-09 | 深圳市华星光电技术有限公司 | Method for compressing image and device |
CN110113609B (en) * | 2019-04-26 | 2020-09-08 | 深圳市华星光电技术有限公司 | Image compression method and device |
WO2020215433A1 (en) * | 2019-04-26 | 2020-10-29 | 深圳市华星光电技术有限公司 | Image compression method and apparatus |
CN110717875A (en) * | 2019-10-18 | 2020-01-21 | 华北理工大学 | High-definition image processing method |
CN110717875B (en) * | 2019-10-18 | 2023-08-08 | 华北理工大学 | High-definition image processing method |
Also Published As
Publication number | Publication date |
---|---|
CN106559668B (en) | 2018-07-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110087092B (en) | Low-bit-rate video coding and decoding method based on image reconstruction convolutional neural network | |
CN107018422B (en) | Still image compression method based on depth convolutional neural networks | |
CN101420614B (en) | Image compression method and device integrating hybrid coding and wordbook coding | |
CN104867165A (en) | Cramping method based on sampling technology under transform domain | |
CN106385584B (en) | The adaptively sampled coding method of distributed video compressed sensing based on spatial correlation | |
JP5022471B2 (en) | Encoding method of wavelet image and corresponding decoding method | |
CN104683811A (en) | Information hiding and extracting method based on integer DCT (Discrete Cosine Transformation) coefficient modulation | |
Mander et al. | An improved image compression-decompression technique using block truncation and wavelets | |
CN105392009A (en) | Low bit rate image coding method based on block self-adaptive sampling and super-resolution reconstruction | |
CN101631243A (en) | Image encoding/decoding method based on wavelet transformation | |
CN106559668B (en) | A kind of low code rate image compression method based on intelligent quantization technology | |
CN104935928B (en) | A kind of efficient image compression method based on spatial domain down-sampling pattern | |
CN111050170A (en) | Image compression system construction method, compression system and method based on GAN | |
CN102333223A (en) | Video data coding method, decoding method, coding system and decoding system | |
CN100515021C (en) | Image quality and constrained code rate based JPEG2000 code rate control method | |
CN111080729B (en) | Training picture compression network construction method and system based on Attention mechanism | |
WO2001050769A1 (en) | Method and apparatus for video compression using multi-state dynamical predictive systems | |
CN105611288B (en) | A kind of low bit rate image sequence coding method based on Constrained interpolation technique | |
CN105872536A (en) | Method for image compression based on dual encoding mode | |
Zhu et al. | An improved SPIHT algorithm based on wavelet coefficient blocks for image coding | |
CN107948644A (en) | A kind of underwater picture compression method and transmission method | |
CN101193285A (en) | Method and device for image compression coding and decoding | |
CN105611289B (en) | Low-resolution image coding method based on intelligent quantization technology | |
CN106131575A (en) | The method for compressing image combined with Chinese remainder theorem based on wavelet transformation | |
CN104469389B (en) | Low bit rate video encoding method and system based on conversion downsampling |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |