CN104683818A - Image compression method based on biorthogonal invariant set multi-wavelets - Google Patents
Image compression method based on biorthogonal invariant set multi-wavelets Download PDFInfo
- Publication number
- CN104683818A CN104683818A CN201510112891.4A CN201510112891A CN104683818A CN 104683818 A CN104683818 A CN 104683818A CN 201510112891 A CN201510112891 A CN 201510112891A CN 104683818 A CN104683818 A CN 104683818A
- Authority
- CN
- China
- Prior art keywords
- image
- inner product
- line
- row
- expansion
- 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
Abstract
The invention discloses an image compression method based on biorthogonal invariant set multi-wavelets. The image compression method comprises the following steps: 1, inputting image data; 2, performing DC level displacement; 3, forming a filter matrix judger; 4, performing biorthogonal invariant set multi-wavelet transform layer number initialization; 5, increasing height and width of an image; 6, performing biorthogonal invariant set multi-wavelet row transform; 7, performing biorthogonal invariant set multi-wavelet line transform; 8, quantifying coefficients; 9, performing arithmetic coding; 10, cutting rate-distortion optimization; 11, organizing code streams. A group of biorthogonal invariant set multi-wavelet filters with symmetry, compact support and orthogonality are adopted, and an inner product method is adopted in biorthogonal invariant set multi-wavelet transform calculation, so that the algorithm complexity is greatly reduced, the energy and entropy concentration degree are high while the sparsity is higher after transform, later compression encoding is facilitated, and blocking processing and parallel acceleration can be more conveniently.
Description
Technical field
The present invention relates to technical field of image processing, further relate to a kind of method for compressing image based on biorthogonal invariant set m ultiwavelet realized in JPEG2000 image compression system in Image Compression field.The present invention can be used for various digital still compressed encoding.
Background technology
Along with development and the application of multimedia and internet, traditional image compression algorithm can not meet the requirement of practical application, and International Standards Organization has formulated the new standard JPEG2000 of still image compression in November, 2000 for this reason.This new standard adopts the rate-distortion optimized truncation built-in code block encryption algorithm (EBCOT) based on wavelet transformation technique, achieves good image compression effect.But JPEG2000 system needs fixed point and floating-point two cover system to realize 5/3 and 9/7 wavelet transformation.
What Xian Electronics Science and Technology University had " realize the fixed point small wave converting method of JPEG2000 image compression " in (number of patent application: 201210148213, publication number: CN102685501B) in its patented technology application discloses a kind of fixed point small wave converting method realizing JPEG2000 image compression.The method introduces BIBO gain control method, use 9/7 small echo BIBO gain dark to determine the bank bit of wavelet transformation median, use 5/3 small echo BIBO gain to determine selection mode and the quantification manner of 9/7 lifting wavelet transform quantization parameter, save JPEG2000 system memory resources and running time.But the weak point that this patented technology exists is, 5/3 small echo adopted due to this patented technology and 9/7 small echo are all single wavelet, every layer of wavelet transformation only can obtain the feature on 4 different directions, after wavelet transformation, the intensity of energy and entropy is not high yet, and need during wavelet transformation that boundary extension is carried out to image and cause Boundary Distortion, be unfavorable for piecemeal process and parallel accelerate.
Liu Wei and Chen Dongli proposes a kind of Image Coding Algorithms method based on multi-wavelet transformation in paper " Image Coding Algorithms based on multi-wavelet transformation is studied " (Zhongshan University's journal (natural science edition) 2011 5 phases the 50th page to 53 pages).The method, by various conventional multi-wavelet bases, in conjunction with EBCOT (the best blocks embedded code block forecast) algorithm, carries out compressed encoding to gray level image.Although this method achieves certain compression effectiveness, but, the deficiency that the method still exists is, the method only gives the simple multi-wavelet bases that 2 take advantage of 2, be difficult to embody symmetry, short supportive, orthogonality and high-order vanishing moment character that m ultiwavelet has simultaneously, and Coding Compression Algorithm computational complexity does not improve with JPEG2000 suitable.
Summary of the invention
The object of the invention is to the deficiency overcoming above-mentioned prior art, provide in a kind of JPEG2000 image compression system based on biorthogonal invariant set multi-wavelet transformation implementation method.The present invention adopt one group have symmetry, tight, orthogonal take advantage of with 4 of high-order vanishing moment characteristic 4 biorthogonal not invariant set multi-wavelet filter matrix, again rank transformation is carried out to the advanced every trade conversion of image, during transformation calculations, adopts the method for inner product.The present invention does not need boundary extension, avoid Boundary Distortion, be conducive to piecemeal process and parallel accelerate, and the present invention's algorithm complex while achieving the effect same with employing method under traditional convolution is but 1/4th of conventional method, the present invention is to energy after image conversion with entropy intensity is high has larger openness in addition, is conducive to follow-up compressed encoding.
For achieving the above object, method of the present invention comprises the steps:
(1) input image data:
Two-dimensional image data to be compressed is inputted in JPEG2000 image compression system;
(2) DC level shift:
DC level shift is carried out to the two-dimensional image data to be compressed of input, obtains the view data after 0 symmetrical DC level shift;
(3) electric-wave filter matrix is formed:
By each biorthogonal invariant set multi-wavelet filter
With
As a line of matrix, composition 4 takes advantage of the electric-wave filter matrix of 4;
(4) biorthogonal invariant set multi-wavelet transformation number of plies initialization;
(5) height and width of expanded images:
(5a) after judging DC level shift, whether the height of image is the multiple of 4, if so, then performs step (5c), otherwise, perform step (5b);
(5b) carry out border symmetric extension to the height of image after DC level shift, after making expansion, the height of image is the multiple of 4;
(5c) judge image after DC level shift wide whether 4 multiple, if so, then perform step (6), otherwise, perform step (5d);
(5d) carry out border symmetric extension to the wide of image, after making expansion, image wide is the multiple of 4, image after being expanded;
(6) biorthogonal invariant set m ultiwavelet line translation:
(6a) respectively the line order k of image and the line order m of line translation image after expansion is initialized as 1;
(6b) four elements of the 1 to 4 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable first element of inner product result as line translation image, k represents the line order number of the rear image of expansion, and m represents the line order number of line translation image; Four elements of the 5 to 8 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable second element of inner product result as line translation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 of rear to electric-wave filter matrix the first row and expansion image row k are carried out inner product, and using m capable the i-th+1 element of inner product result as line translation image, i presentation video element numbers, the span of i is
w represents the width of the rear image of expansion;
(6c) four elements of the 1 to 4 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix second row and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6d) four elements of the 1 to 4 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix the third line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6e) four elements of the 1 to 4 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the expression line order number of k image, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix fourth line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6f) respectively the line order k of image and the line order m of line translation image after expansion is added 1;
(6g) after judging to add the expansion after 1, whether the line order k of image equals to expand the height of rear image, if so, then obtains line translation image, performs step (7), otherwise, perform step (6b);
(7) biorthogonal invariant set m ultiwavelet rank transformation:
(7a) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is initialized as 1;
(7b) four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 1 to 4 carry out inner product, using inner product result r row first element as rank transformation image, and the row sequence number of n presentation video; Four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 5 to 8 carry out inner product, using inner product result r row second element as rank transformation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the first row and line translation image n-th arranged carry out inner product, and using inner product result r row the i-th+1 element as rank transformation image, i presentation video element numbers, the span of i is
l represents the height of the rear image of expansion;
(7c) four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix second row and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7d) four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the third line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7e) four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix fourth line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7f) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is added 1;
(7g) after judging to add the expansion after 1, whether the row sequence number n of image equals to expand the width of rear image, if so, then obtains biorthogonal invariant set multi-wavelet transformation image, performs step (8), otherwise, perform step (7b);
(8) more new images:
Orthogonally-persistent collection multi-wavelet transformation image is existed
the height of scope and
wide part in scope, as image after DC level shift;
(9) the conversion number of plies of biorthogonal invariant set m ultiwavelet is subtracted 1;
(10) after judging to subtract 1, whether the biorthogonal invariant set multi-wavelet transformation number of plies is zero, and after being if so, expanded, the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient of image, then perform step (11), otherwise, perform step (5);
(11) quantization parameter:
By the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient quantization of image after expansion;
(12) arithmetic coding:
Utilize the coefficient after wavelet transformation being processed based on contextual bit plane arithmetic coding module of JPEG2000 image compression system Plays, obtain the code stream of arithmetic coding;
(13) rate-distortion optimization:
The rate-distortion optimized truncation module of JPEG2000 image compression system Plays is utilized to carry out rate-distortion optimized truncation to the code stream of arithmetic coding, record intercept point information;
(14) tissue code stream:
The code stream organization module of JPEG2000 image compression system Plays uses intercept point information, carries out code stream organization, obtain the compressed bit stream of JPEG2000 to the code stream of arithmetic coding.
Compared with prior art, tool has the following advantages in the present invention:
First, owing to present invention employs the biorthogonal invariant set multi-wavelet filter matrix that 4 take advantage of 4, each biorthogonal invariant set m ultiwavelet filtering transformation can obtain the feature on 16 different directions, overcome single wavelet every layer wavelet transformation in prior art and only can obtain the shortcoming of the feature on 4 different directions, make the present invention have the advantage of more direction presentation video.
Second, the method of inner product is have employed during feature and transformation calculations due to biorthogonal invariant set multi-wavelet filter matrix of the present invention, overcome the shortcoming that the intensity of single wavelet transformation energy and entropy in prior art is not high, make the present invention have the advantage of better compression performance.
3rd, the method of inner product is have employed during feature and transformation calculations due to biorthogonal invariant set multi-wavelet filter matrix of the present invention, need that boundary extension is carried out to image when overcoming single wavelet conversion in prior art and cause Boundary Distortion, be unfavorable for the shortcoming of piecemeal process and parallel accelerate, make the present invention have the advantage realizing compression fast.
4th, because the present invention 4 takes advantage of the method that have employed inner product when the feature of the biorthogonal invariant set multi-wavelet filter matrix of 4 and transformation calculations, to overcome in prior art 2 and take advantage of the shortcoming that 2 simple multi-wavelet bases Coding Compression Algorithm computational complexities are high, make the present invention have algorithm complex to reduce, save operation time, realize the advantage of compression fast.
Accompanying drawing explanation
Fig. 1 is flow chart of the present invention.
Embodiment
Below in conjunction with accompanying drawing 1, performing step of the present invention is described in detail.
Step 1. input image data.
In JPEG2000 image compression system, input two-dimensional image data to be compressed, pixel value adopts 1 ~ 16 bit.
Step 2.DC level shift.
DC level shift is carried out to the two-dimensional image data to be compressed of input, obtains the view data after 0 symmetrical DC level shift.
Step 3. forms electric-wave filter matrix.
By each biorthogonal invariant set multi-wavelet filter
With
As a line of matrix, composition 4 takes advantage of the electric-wave filter matrix of 4 to be shown below:
Wherein, H
4 × 4the biorthogonal invariant set multi-wavelet filter matrix of 4 is taken advantage of in expression 4.
Biorthogonal invariant set multi-wavelet filter H
4 × 4it is the biorthogonal invariant set m ultiwavelet theory had from imitative map feature set up in article " Reconstruction and Decomposition Algorithms for Biorthogonal Multiwavelets " according to Charles A.Micchelli and Yuesheng Xu, there is the Delta Region of self affine for support Interval, with the constant function on it for scaling function, construct one group and there is symmetry, tight, orthogonal biorthogonal not invariant set multi-wavelet filter.This electric-wave filter matrix has the shortest support length, does not have support lap, and energy Accurate Reconstruction after making wavelet decomposition, non-boundary distortion effect, avoids boundary extension, just with piecemeal process and parallel accelerate when applying.
The initialization of the step 4. biorthogonal invariant set multi-wavelet transformation number of plies.
Biorthogonal invariant set multi-wavelet transformation is generally no more than 3 layers, usually gets 1 layer or 2 layers, therefore generally the biorthogonal invariant set multi-wavelet transformation number of plies is initialized as 1 or 2.
The height and width of step 5. expanded images.
(5a) after judging DC level shift, whether the height of image is the multiple of 4, if so, then performs step (5c), otherwise, perform step (5b).
(5b) carry out border symmetric extension to the height of image after DC level shift, after making expansion, the height of image is the multiple of 4.
(5c) judge image after DC level shift wide whether 4 multiple, if so, then perform step 6, otherwise, perform step (5d).
(5d) carry out border symmetric extension to the wide of image, after making expansion, image wide is the multiple of 4, image after being expanded.
The multiple being 4 by the height and width symmetric extension of image after DC level shift refers to, the height and width of the image after expansion can be divided exactly by 4, and the line number of expansion or columns are in the scope of 1 to 3.
The line translation of step 6. biorthogonal invariant set m ultiwavelet.
(6a) respectively the line order k of image and the line order m of line translation image after expansion is initialized as 1.
(6b) four elements of the 1 to 4 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable first element of inner product result as line translation image, k represents the line order number of the rear image of expansion, and m represents the line order number of line translation image; Four elements of the 5 to 8 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable second element of inner product result as line translation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 of rear to electric-wave filter matrix the first row and expansion image row k are carried out inner product, and using m capable the i-th+1 element of inner product result as line translation image, i presentation video element numbers, the span of i is
w represents the width of the rear image of expansion.
(6c) four elements of the 1 to 4 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix second row and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6d) four elements of the 1 to 4 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix the third line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6e) four elements of the 1 to 4 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the expression line order number of k image, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix fourth line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6f) respectively the line order k of image and the line order m of line translation image after expansion is added 1.
(6g) after judging to add the expansion after 1, whether the line order k of image equals to expand the height of rear image, if so, then obtains line translation image, performs step 7, otherwise, perform step (6b).
Step 7. biorthogonal invariant set m ultiwavelet rank transformation.
(7a) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is initialized as 1.
(7b) four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 1 to 4 carry out inner product, using inner product result r row first element as rank transformation image, and the row sequence number of n presentation video; Four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 5 to 8 carry out inner product, using inner product result r row second element as rank transformation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the first row and line translation image n-th arranged carry out inner product, and using inner product result r row the i-th+1 element as rank transformation image, i presentation video element numbers, the span of i is
l represents the height of the rear image of expansion.
(7c) four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix second row and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7d) four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the third line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7e) four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix fourth line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7f) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is added 1.
(7g) after judging to add the expansion after 1, whether the row sequence number n of image equals to expand the width of rear image, if so, then obtains biorthogonal invariant set multi-wavelet transformation image, performs step 8, otherwise, perform step (7b).
Step 8. is new images more.
Orthogonally-persistent collection multi-wavelet transformation image is existed
the height of scope and
wide part in scope, as image after DC level shift.
The conversion number of plies of biorthogonal invariant set m ultiwavelet is subtracted 1 by step 9..
After step 10. judges to subtract 1, whether the biorthogonal invariant set multi-wavelet transformation number of plies is zero, and the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient of the image after being if so, expanded, then perform step 11, otherwise, perform step 5.
The feature that this biorthogonal invariant set multi-wavelet transformation process make use of biorthogonal invariant set multi-wavelet filter have employed the method for inner product, thus it is same with employing method under traditional convolution to obtain transform effect, algorithm complex is but the advantage of 1/4th of conventional method, and energy and entropy intensity is high has larger openness after conversion, be conducive to follow-up compressed encoding.
Step 11. quantization parameter.
By the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient quantization of the image after expansion, during quantification, utilize the quantizing process of the international standard ISO/IEC 15444-1:2000 of JPEG2000 image compression system.
Due to the feature on 16 different directions of the original image of every layer of biorthogonal invariant set multi-wavelet transformation, and every layer every layer 9/7 wavelet transformation can only obtain the feature on 4 different directions of original image, in order to the quantizing process of JPEG2000 image compression system Plays can be utilized, can biorthogonal invariant set multi-wavelet transformation and 9/7 wavelet transformation do one corresponding.To during correspondence, biorthogonal invariant set multi-wavelet transformation is mapped just passable with the low frequency part of 9/7 wavelet transformation, correspondence can be carried out as required in other directions.
Step 12. arithmetic coding.
Utilize the coefficient after wavelet transformation being processed based on contextual bit plane arithmetic coding module of JPEG2000 image compression system Plays, obtain the code stream of arithmetic coding.
Step 13. rate-distortion optimization.
The rate-distortion optimized truncation module of JPEG2000 image compression system Plays is utilized to carry out rate-distortion optimized truncation to the code stream of arithmetic coding, record intercept point information.
Step 14. organizes code stream.
The code stream organization module of JPEG2000 image compression system Plays uses intercept point information, carries out code stream organization, obtain the compressed bit stream of JPEG2000 to the code stream of arithmetic coding.
Due to the compressed bit stream of JPEG2000 obtained, what wavelet transformation part adopted is biorthogonal invariant set multi-wavelet transformation, so when the compressed bit stream of this JPEG2000 carries out decompress(ion), inverse wavelet transform part also needs to adopt biorthogonal invariant set m ultiwavelet inverse transformation accordingly.
Claims (4)
1., based on a method for compressing image for biorthogonal invariant set m ultiwavelet, comprise the steps:
(1) input image data:
Two-dimensional image data to be compressed is inputted in JPEG2000 image compression system;
(2) DC level shift:
DC level shift is carried out to the two-dimensional image data to be compressed of input, obtains the view data after 0 symmetrical DC level shift;
(3) electric-wave filter matrix is formed:
By each biorthogonal invariant set multi-wavelet filter
With
As a line of matrix, composition 4 takes advantage of the electric-wave filter matrix of 4;
(4) biorthogonal invariant set multi-wavelet transformation number of plies initialization;
(5) height and width of expanded images:
(5a) after judging DC level shift, whether the height of image is the multiple of 4, if so, then performs step (5c), otherwise, perform step (5b);
(5b) carry out border symmetric extension to the height of image after DC level shift, after making expansion, the height of image is the multiple of 4;
(5c) judge image after DC level shift wide whether 4 multiple, if so, then perform step (6), otherwise, perform step (5d);
(5d) carry out border symmetric extension to the wide of image, after making expansion, image wide is the multiple of 4, image after being expanded;
(6) biorthogonal invariant set m ultiwavelet line translation:
(6a) respectively the line order k of image and the line order m of line translation image after expansion is initialized as 1;
(6b) four elements of the 1 to 4 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable first element of inner product result as line translation image, k represents the line order number of the rear image of expansion, and m represents the line order number of line translation image; Four elements of the 5 to 8 of rear to electric-wave filter matrix the first row and expansion image row k the are carried out inner product, using m capable second element of inner product result as line translation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 of rear to electric-wave filter matrix the first row and expansion image row k are carried out inner product, and using m capable the i-th+1 element of inner product result as line translation image, i presentation video element numbers, the span of i is
w represents the width of the rear image of expansion;
(6c) four elements of the 1 to 4 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix second row and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix second row and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6d) four elements of the 1 to 4 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the line order number of k presentation video, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix the third line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix the third line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6e) four elements of the 1 to 4 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element, the expression line order number of k image, m represents the line order number of line translation image, and W represents the width of the rear image of expansion; Four elements of the 5 to 8 of image row k after electric-wave filter matrix fourth line and expansion the are carried out inner product, using inner product result as the m of line translation image capable the
individual element; The rest may be inferred, and four elements of 4th × i+1 to 4 × i+4 of image row k after electric-wave filter matrix fourth line and expansion are carried out inner product, using inner product result as the m of line translation image capable the
individual element, i presentation video element numbers, the span of i is
(6f) respectively the line order k of image and the line order m of line translation image after expansion is added 1;
(6g) after judging to add the expansion after 1, whether the line order k of image equals to expand the height of rear image, if so, then obtains line translation image, performs step (7), otherwise, perform step (6b);
(7) biorthogonal invariant set m ultiwavelet rank transformation:
(7a) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is initialized as 1;
(7b) four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 1 to 4 carry out inner product, using inner product result r row first element as rank transformation image, and the row sequence number of n presentation video; Four elements of the electric-wave filter matrix the first row and line translation image n-th are arranged the 5 to 8 carry out inner product, using inner product result r row second element as rank transformation image; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the first row and line translation image n-th arranged carry out inner product, and using inner product result r row the i-th+1 element as rank transformation image, i presentation video element numbers, the span of i is
l represents the height of the rear image of expansion;
(7c) four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix second row and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix second row and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7d) four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix the third line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix the third line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7e) four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 1 to 4 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, the row sequence number of n presentation video, L represents the height of the rear image of expansion; Four elements of the electric-wave filter matrix fourth line and line translation image n-th are arranged the 5 to 8 carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element; The rest may be inferred, and four elements of the 4th × i+1 to 4 × i+4 electric-wave filter matrix fourth line and line translation image n-th arranged carry out inner product, and inner product result is arranged as the r of rank transformation image
individual element, i presentation video element numbers, the span of i is
(7f) respectively the row sequence number n of image and the row sequence number r of rank transformation image after expansion is added 1;
(7g) after judging to add the expansion after 1, whether the row sequence number n of image equals to expand the width of rear image, if so, then obtains biorthogonal invariant set multi-wavelet transformation image, performs step (8), otherwise, perform step (7b);
(8) more new images:
Orthogonally-persistent collection multi-wavelet transformation image is existed
the height of scope and
wide part in scope, as image after DC level shift;
(9) the conversion number of plies of biorthogonal invariant set m ultiwavelet is subtracted 1;
(10) after judging to subtract 1, whether the biorthogonal invariant set multi-wavelet transformation number of plies is zero, and after being if so, expanded, the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient of image, then perform step (11), otherwise, perform step (5);
(11) quantization parameter:
By the plurality of layers of double orthogonally-persistent collection multi-wavelet transformation coefficient quantization of image after expansion;
(12) arithmetic coding:
Utilize the coefficient after wavelet transformation being processed based on contextual bit plane arithmetic coding module of JPEG2000 image compression system Plays, obtain the code stream of arithmetic coding;
(13) rate-distortion optimization:
The rate-distortion optimized truncation module of JPEG2000 image compression system Plays is utilized to carry out rate-distortion optimized truncation to the code stream of arithmetic coding, record intercept point information;
(14) tissue code stream:
The code stream organization module of JPEG2000 image compression system Plays uses intercept point information, carries out code stream organization, obtain the compressed bit stream of JPEG2000 to the code stream of arithmetic coding.
2. the method for compressing image based on biorthogonal invariant set m ultiwavelet according to claim 1, is characterized in that: the two-dimensional image data to be compressed described in step (1) adopts 1 ~ 16 bit.
3. the method for compressing image based on biorthogonal invariant set m ultiwavelet according to claim 1, it is characterized in that: the multiple being 4 by the height and width symmetric extension of image after DC level shift described in step (5) refers to, after making expansion, the height and width of image can be divided exactly by 4, and the line number of expansion or columns are in the scope of 1 to 3.
4. the method for compressing image based on biorthogonal invariant set m ultiwavelet according to claim 1, it is characterized in that: the multi-wavelet transformation of plurality of layers of double orthogonally-persistent collection described in step (11) coefficient quantization refers to, utilizes the quantizing process of the international standard ISO/IEC 15444-1:2000 of JPEG2000 image compression system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510112891.4A CN104683818B (en) | 2015-03-15 | 2015-03-15 | Method for compressing image based on biorthogonal invariant set m ultiwavelet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510112891.4A CN104683818B (en) | 2015-03-15 | 2015-03-15 | Method for compressing image based on biorthogonal invariant set m ultiwavelet |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104683818A true CN104683818A (en) | 2015-06-03 |
| CN104683818B CN104683818B (en) | 2017-11-21 |
Family
ID=53318298
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510112891.4A Active CN104683818B (en) | 2015-03-15 | 2015-03-15 | Method for compressing image based on biorthogonal invariant set m ultiwavelet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104683818B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106056158A (en) * | 2016-06-03 | 2016-10-26 | 西安电子科技大学 | Template image global searching method based on mask matrix and fast Fourier transform |
| CN106664408A (en) * | 2014-06-04 | 2017-05-10 | 简·克劳德·科林 | Adaptive precision and quantification of a wavelet transformed matrix |
| CN107071482A (en) * | 2017-03-31 | 2017-08-18 | 东南大学 | A kind of compression of images and reconstructing method based on multilayer discrete wavelet transform coefficients |
| CN111885390A (en) * | 2020-07-30 | 2020-11-03 | 河南大学 | High spectral image compression method based on fractal multi-wavelet |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070030177A1 (en) * | 2003-09-18 | 2007-02-08 | Monro Donald M | Data compression |
| CN102685501A (en) * | 2012-05-14 | 2012-09-19 | 西安电子科技大学 | Fixed-point wavelet transform method for joint photographic experts group 2000 (JPEG2000) image compression |
-
2015
- 2015-03-15 CN CN201510112891.4A patent/CN104683818B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070030177A1 (en) * | 2003-09-18 | 2007-02-08 | Monro Donald M | Data compression |
| CN102685501A (en) * | 2012-05-14 | 2012-09-19 | 西安电子科技大学 | Fixed-point wavelet transform method for joint photographic experts group 2000 (JPEG2000) image compression |
Non-Patent Citations (1)
| Title |
|---|
| 张静 等: "JPEG2000 算法中基于有界输入有界输出增益控制的小波变换定点实现技术", 《电子与信息学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106664408A (en) * | 2014-06-04 | 2017-05-10 | 简·克劳德·科林 | Adaptive precision and quantification of a wavelet transformed matrix |
| CN106664408B (en) * | 2014-06-04 | 2019-08-20 | 简·克劳德·科林 | A method of compression digital picture |
| CN106056158A (en) * | 2016-06-03 | 2016-10-26 | 西安电子科技大学 | Template image global searching method based on mask matrix and fast Fourier transform |
| CN106056158B (en) * | 2016-06-03 | 2019-06-21 | 西安电子科技大学 | Template image global search method based on exposure mask matrix and Fast Fourier Transform |
| CN107071482A (en) * | 2017-03-31 | 2017-08-18 | 东南大学 | A kind of compression of images and reconstructing method based on multilayer discrete wavelet transform coefficients |
| CN107071482B (en) * | 2017-03-31 | 2019-12-10 | 东南大学 | Image compression and reconstruction method based on multilayer discrete wavelet transform coefficients |
| CN111885390A (en) * | 2020-07-30 | 2020-11-03 | 河南大学 | High spectral image compression method based on fractal multi-wavelet |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104683818B (en) | 2017-11-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102123278B (en) | Signal source encoding and decoding method based on distributed compressive sensing technology | |
| CN104331913A (en) | Polarized SAR polarization method based on sparse K-SVD (Singular Value Decomposition) | |
| CN104574336A (en) | Super-resolution image reconstruction system based on self-adaptation submodel dictionary choice | |
| Yadav et al. | A review on image compression techniques | |
| CN103905831B (en) | Strip wave transform image compression method based on graphics processor | |
| CN104427349A (en) | Bayer image compression method | |
| CN105847800A (en) | Image compression method based on all phase discrete sine double orthogonal transformation and system | |
| CN104683818A (en) | Image compression method based on biorthogonal invariant set multi-wavelets | |
| Biswas et al. | A high-precision low-area unified architecture for lossy and lossless 3D multi-level discrete wavelet transform | |
| CN113962882B (en) | JPEG image compression artifact eliminating method based on controllable pyramid wavelet network | |
| CN102982563B (en) | A kind of method for compressing image based on optical profile wave conversion | |
| CN107948644B (en) | Underwater image compression method and transmission method | |
| CN115358929A (en) | Compressed image super-resolution method, image compression method and system | |
| Turcza | Entropy encoder for low-power low-resources high-quality CFA image compression | |
| CN104869426A (en) | JPEG coding method lowering image diamond effect under low compression code rate | |
| Zhang et al. | Memory-efficient high-speed VLSI implementation of multi-level discrete wavelet transform | |
| Divakara et al. | High speed modular systolic array-based DTCWT with parallel processing architecture for 2D image transformation on FPGA | |
| CN109255770B (en) | Image transform domain down-sampling method | |
| Kountchev et al. | Decorrelation of multispectral images, based on hierarchical adaptive PCA | |
| Sendamarai et al. | FPGA implementation of combined compression & denoising scheme for remote sensing images | |
| Rani et al. | Improving accuracy of deep learning-based compression techniques by introducing perceptual loss in industrial IoT | |
| Rajakumar et al. | Lossy image compression using multiwavelet transform for wireless transmission | |
| Mehta et al. | Image Compression using DCT and DWT-Technique | |
| Rajeshwari et al. | DWT based Multimedia Compression | |
| CN115601242B (en) | Lightweight image super-resolution reconstruction method suitable for hardware deployment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |