CN105376578A - Image compression method and device - Google Patents

Abstract

The invention discloses an image compression method and device. The method comprise that original data of an image is preprocessed; WALSH transformation is carried out on image data which is obtained via preprocessing; and image compression is carried out on the transformed image data in an entropy coding algorithm. According to the method and device, redundant information in the image can be effectively removed, and the coding efficiency as well as the compression ratio is improved.

Description

Method for compressing image and device

Technical field

The embodiment of the present invention relates to technical field of image processing, particularly relates to a kind of method for compressing image and device.

Background technology

Along with life enters the information age, people by more dependence equipment obtaining information from the Internet, thus need a large amount of storages, record and transmission information.But a large amount of information represents with digital form, store, transmits, in these digitized information, image occupies great space.In order to enable the information data of image effectively store, transmit and utilize, the size of necessary packed data.Image can be considered as a huge information aggregate, describe that will to comprise in these information be amount of information and information redundancy etc.Image Data Compression Technique is exactly how research utilizes the redundancy of view data to reduce the method for image data amount.

As shown in Figure 1, for the general step of image compression, mainly comprise coding stage and decode phase, wherein coding stage comprises: the first step carries out mapping transformation to view data, adopt its object be in this way exactly in order to removal of images inside most correlation, its redundancy is reduced greatly.Through mapping phase, the data characteristics of original image will be changed, and advantageously encodes in picture compression.Second step is quantizing process, and essence is that the parameter removing redundant information formation is entered the entropy code stage.3rd step is statistical coding i.e. entropy code, can Using statistics coded treatment symbol after completing quantification, and this also can reduce the unnecessary data volume of image, and compression is completed, and statistical coding conventional at present comprises Huffman coding etc.The decode procedure that the flow process reverse operating of said process is exactly compression, mainly comprise: encoding stream is decoded, then carry out inverse quantization, finally carry out anti-mapping transformation.

And in prior art, in existing image compression process, the mapping transformation of employing is dct transform, and dct transform effectively can not remove the redundant information in image, cause in next code process and contain more redundant information, reduce code efficiency and compression ratio.

Summary of the invention

The embodiment of the present invention provides a kind of method for compressing image and device, effectively to remove the redundant information in image, improves code efficiency and compression ratio.

First aspect, embodiments provides a kind of method for compressing image, comprising:

Preliminary treatment is carried out to original image data;

The view data obtained for preliminary treatment carries out Walsh WALSH conversion;

Entropy code algorithm is adopted to carry out image compression for the view data after conversion.

Second aspect, the embodiment of the present invention also provides a kind of image compressing device, comprising:

Pretreatment module, for carrying out preliminary treatment to original image data;

Conversion module, the view data for obtaining for preliminary treatment carries out Walsh WALSH conversion;

Compressed encoding module, for adopting entropy code algorithm to carry out image compression for the view data after conversion.

The embodiment of the present invention is by carrying out preliminary treatment to original image data; The view data obtained for preliminary treatment carries out Walsh WALSH conversion; Entropy code algorithm is adopted to carry out image compression for the view data after conversion.The embodiment of the present invention effectively can remove the redundant information in image by use WALSH, improve code efficiency and compression ratio.

Accompanying drawing explanation

The schematic flow sheet of the method for compressing image that Fig. 1 provides for prior art;

The schematic flow sheet of the method for compressing image that Fig. 2 A provides for the embodiment of the present invention one;

The schematic flow sheet of the image block in the method for compressing image that Fig. 2 B provides for the embodiment of the present invention one;

The scanning process schematic diagram of the zigzag in the method for compressing image that Fig. 2 C provides for the embodiment of the present invention one;

The structural representation of the image compressing device that Fig. 3 provides for the embodiment of the present invention two.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.Be understandable that, specific embodiment described herein is only for explaining the present invention, but not limitation of the invention.It also should be noted that, for convenience of description, illustrate only part related to the present invention in accompanying drawing but not entire infrastructure.

The executive agent of the method for compressing image that the embodiment of the present invention provides, can be the image compressing device that the embodiment of the present invention provides, or be integrated with the terminal equipment of described image compressing device (such as, smart mobile phone, panel computer etc.), this image compressing device can adopt hardware or software simulating.

Embodiment one

The schematic flow sheet of the method for compressing image that Fig. 2 A provides for the embodiment of the present invention one, as shown in Figure 2, specifically comprises the steps:

Step 11, preliminary treatment is carried out to original image data;

Wherein, described preliminary treatment comprises the operation such as image format conversion and/or fragmental image processing.

Step 12, the view data obtained for preliminary treatment carry out Walsh WALSH conversion;

Wherein, it is a kind of orthogonal transform that WALSH converts, and it only comprises two numerical value that value is "+1 " and "-1 " and forms whole friendship function base.Although conventional DCT has many advantages, its inferior position also clearly, first quantizing process more complicated, next needs the computational process by multistep complexity, and hardware implementing get up neither be very convenient.And the function base of WALSH conversion is two-value orthogonal basis, corresponding with two states of Digital Logic, so be more suitable for computer disposal.WALSH conversion relatively reduces memory space and arithmetic speed is obtained and promotes.

Concrete, WALSH function is by incomplete Rademacher function in addition completion, and define one group of complete orthogonal matrix function, its value can only be+1 and-1.During normalization, function is defined as in interval [0,1]:

$wal(n,t)=\underset{m=0}{\overset{p-1}{\Π}}{\[r(m+1,t)\]}^{g_m}$

Wherein, n=0,1,2 ..., p is the binary coding figure place of n, g _mfor the m bit value of the Gray code of n.

Wherein, the exponential form of WALSH conversion is as follows:

$wal(n,t)=\underset{m=0}{\overset{p-1}{\Π}}{(-1)}^{t_m{g}_m}$

Visible, this WALSH convert exponential function definition be+1 or-1 continued product.The positive and negative change number of times of each waveform of front 8 waveforms of WALSH conversion is n.If interval to be divided into N decile, to the waveform sampling in each interval, a discrete functional value matrix can be obtained, this matrix be 8 × 8 matrix as follows:

$W_8=\left[\begin{array}{cccccccc}1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& -1& -1& -1& -1\\ 1& 1& -1& -1& -1& -1& 1& 1\\ 1& 1& -1& -1& 1& 1& -1& -1\\ 1& -1& -1& 1& 1& -1& -1& 1\\ 1& -1& -1& 1& -1& 1& 1& -1\\ 1& -1& 1& -1& -1& 1& -1& 1\\ 1& -1& 1& -1& 1& -1& 1& -1\end{array}\right]$

Represent 8 sub-samplings of DISCRETE W ALSH function in above formula, T represents sampling sequence number, be from left to right the 0th, 1..... the 7th time.

For example, picture signal can be carried out data compression, by lower column signal by two-dimentional WALSH conversion $X=\left[\begin{array}{cccc}0& 1& 1& 0\\ 0& 1& 1& 0\\ 0& 1& 1& 0\\ 0& 1& 1& 0\end{array}\right]$ Carry out the WALSH conversion of 4 sub-samplings $W_4=\left[\begin{array}{cccc}1& 1& 1& 1\\ 1& 1& -1& -1\\ 1& -1& -1& 1\\ 1& -1& 1& -1\end{array}\right],$ And observed result Y:

$Y=\frac{1}{16}{W}_4{\mathrm{XW}}_4=\left[\begin{array}{cccc}\frac{1}{2}& 0& -\frac{1}{2}& 0\\ 0& 0& 0& 0\\ 0& 0& 0& 0\\ 0& 0& 0& 0\end{array}\right]$

Can draw to draw a conclusion from result, only in the part in the upper left corner, some information are remained for initial two-dimensional matrix data after conversion, and the information of remaining most position be all 0.If transmission is the matrix after conversion instead of sampled signal at the beginning, and can transmit the Partial Elements in the upper left corner, the object of packed data just can be reached completely.

Therefore, can reach the expection object of data compression with WALSH conversion, be because eliminate the redundant dependencies of original data message in its process.And the coefficient of matrix representative low frequency after the conversion obtained and flip-flop is at upper left hand corner section, the correspondence image conversion of these coefficients is slowly, corresponding, and the conversion coefficient of instruction radio-frequency component is at right lower quadrant, and in their WALSH image, conversion rate is fast.When encoding to conversion coefficient, the method completing data compression can use region filter method and different accuracy quantization method, and the different quantification numbers of plies is got in different region, and the size of the number of plies depends on the numerical value of the signal energy that this region has.

Step 13, entropy code algorithm is adopted to carry out image compression for the view data after conversion.

Wherein, entropy code algorithm can comprise Huffman encryption algorithm and Arithmetic Coding algorithm.

The present embodiment is by carrying out preliminary treatment to original image data; The view data obtained for preliminary treatment carries out Walsh WALSH conversion; Entropy code algorithm is adopted to carry out image compression for the view data after conversion.The present embodiment effectively can remove the redundant information in image by use WALSH, improve code efficiency and compression ratio.

Exemplary, on the basis of above-described embodiment, preliminary treatment is carried out to original image data and comprises:

The original image data of RGB rgb format is converted to the view data of yuv format;

Piecemeal process is carried out to the view data of yuv format.

Wherein, preliminary treatment is by the aberration rgb signal in image, converts the YUV color difference signal required for quantization encoding to, the step such as change quantization coding after the composition minimum code unit MCU exporting 68 × 8 through a series of computing just can carry out.Usual pretreatment stage mainly comprises following two aspects:

First aspect: color conversion

Under normal conditions, the color space of the static images of compression is identified by RGB.Because in JEPG, the component of color space is made up of YCbCr, and another kind of saying is YUV, and namely U represents Cb, and V represents that Cr.Y represents the brightness of static images, and UV all represents the chromatic component of static images.Therefore, in the process of compressed picture, first need static images to change these two kinds different color space component, be converted to yuv format by rgb format, in decompression, need to allow yuv format be reduced into rgb format original image.Its conversion formula is:

$\left[\begin{array}{c}Y\\ U\\ V\end{array}\right]=\left[\begin{array}{ccc}0.2990& 0.5870& 0.1140\\ -0.1684& -0.3316& 0.5000\\ 0.5000& -0.4187& -0.0813\end{array}\right]\left[\begin{array}{c}R\\ G\\ B\end{array}\right]$

Analysis piece image be the rgb signal that first can obtain this image, in order to carry out image compression, first will convert rgb signal to YUV signal.The rgb signal of source images is passed in calculation block ConvertRgbYuv, because the division proportion of regulation aberration is 4:1:1, so the value of regulation HorizontalSize and VerticalSize is 16, the rgb signal of 16 × 16 dress is changed into luminance component and the color difference components of 16 × 16.

Second aspect: sampling and piecemeal

Next the YUV signal obtained after conversion is sampled, now component YUV is regarded respectively as three matrixes of 16 × 16, because regulation aberration is divided into 4:1:1, so luminance component needs luminance component to need to be divided into the matrix of 48 × 8, put into YUV_1 to YUV_4 respectively.For chromatic component UV, only need get the value of even number line in matrix and be respectively put in YUV_5 and YUV_6.Form a maximum arithmetic element by this matrix of 68 × 8, sampling partitioned mode as shown in Figure 2 B.

Exemplary, on the basis of above-described embodiment, adopt entropy code algorithm to carry out image compression for the view data after conversion and comprise:

Quantification treatment is carried out to the view data after conversion;

Zigzag scanning is carried out to the view data after quantizing, obtains corresponding image vector matrix;

Arithmetic Coding algorithm is adopted to encode to image vector matrix.

Exemplary, adopt Arithmetic Coding algorithm to carry out coding to image vector and comprise:

The probability that in described image vector matrix, each symbol is corresponding is calculated by probability Estimation;

According to each symbol of described probability stamps;

Determine between code area according to the probability stamps of symbol;

Output bit flow is determined according between described code area.

Wherein, described coding interval range is 2 ⁸~ 2 ⁹.

Concrete, WALSH transform and quantization process is the significant process of image compression, is also the essential step of removal of images garbage.It is quantizing process after pre-treatment step, quantification can make the data volume of coding decline, and it is converted to the centrifugal pump of finite data, deletes and changes unconspicuous information at human eye to observation and judgement, namely belong to the data message that human eye can not be perceived easily, data are effectively compressed.JPEG adopts linear homogeneous quantizer, the length and width of quantization table are 8 × 8, need the corresponding quantization step of definition 64 coefficients, each coefficient is one to one, but concrete step-length is unfixing, quantization step is determined by quantizing factor scale simultaneously, its controlled imaged compression quality and compression ratio.

WALSH coefficient carries out layout again after quantizing, and the array of 8 × 8 is stored in a linear fashion, and in order to add the number length of continuous coefficients for " 0 ".The order rearranged scans with the track of advancing of zigzag, as shown in Figure 2 C, such chronological order just can change the array of 8 × 8 into the single line vector matrix of 1 × 64, also have a benefit to be the upper left corner can be represented the front end that low-frequency part is placed on vector, then carrying out ensuing coded treatment and comprise the process of DC coefficient and exchange coefficient.

This coding structure be input as Bin, Bin represents character to be encoded.Before interal separation, coding siding-to-siding block length R and lower limit L to be determined and update probability state constantly by probability Estimation large probability character MPS and small probability character LPS.During interval division, MPS is interval front, and LPS is interval rear.Then judge that the type of current sign is determined between next code area.When inputting MPS, MPS is as between next code area, and lower limit L is constant.When inputting LPS, LPS is as between next code area, and interval limit L will increase between the code area of MPS simultaneously.2 are remained in the interval solemnity of new coding after division ⁸~ 2 ⁹in scope, when being less than 2 ⁸time, just need through reforming.Reforming process can likely export one or more " 0 ", and " 1 " bit, as the output of arithmetic coding, is finally the form with bit stream.

The various embodiments described above are equally by carrying out preliminary treatment to original image data; The view data obtained for preliminary treatment carries out Walsh WALSH conversion; Entropy code algorithm is adopted to carry out image compression for the view data after conversion.The same redundant information can effectively removed by use WALSH in image, improves code efficiency and compression ratio.

The various embodiments described above receive intelligent terminal by control terminal equally and send the characteristic message comprising facility information, and the facility information of described intelligent terminal is got by resolving described characteristic message, the intelligent terminal only conformed to a predetermined condition to described facility information sends network connection information, and the intelligent terminal conformed to a predetermined condition to make facility information is linked in intelligent terminal system according to described network connection information.The various embodiments described above equally only allow the intelligent terminal conformed to a predetermined condition to be linked in intelligent terminal system, thus can add in intelligent terminal system by specific intelligent terminal targetedly, reduce the potential safety hazard of intelligent terminal system.

Embodiment two

The structural representation of the image compressing device that Fig. 3 provides for the embodiment of the present invention three, as shown in Figure 3, specifically comprises: pretreatment module 21, conversion module 22 and compressed encoding module 23;

Described pretreatment module 21 is for carrying out preliminary treatment to original image data;

Described conversion module 22 carries out Walsh WALSH conversion for the view data obtained for preliminary treatment;

Described compressed encoding module 23 is for adopting entropy code algorithm to carry out image compression for the view data after conversion.

Image compressing device described in the present embodiment is for performing the method for compressing image described in the various embodiments described above, and the technique effect of its know-why and generation is similar, is not repeated here.

Exemplary, on the basis of above-described embodiment, described pretreatment module 21 for:

The original image data of RGB rgb format is converted to the view data of yuv format; Piecemeal process is carried out to the view data of yuv format.

Exemplary, on the basis of above-described embodiment, described compressed encoding module 23 comprises: quantification treatment unit 231, scanning element 232 and arithmetic coding unit 233;

Described quantification treatment unit 231 is for carrying out quantification treatment to the view data after conversion;

Described scanning element 232, for carrying out zigzag scanning to the view data after quantification, obtains corresponding image vector matrix;

Described arithmetic coding unit 233 is encoded to image vector matrix for adopting Arithmetic Coding algorithm.

Exemplary, described arithmetic coding unit 233 specifically for:

The probability that in described image vector matrix, each symbol is corresponding is calculated by probability Estimation; According to each symbol of described probability stamps; Determine between code area according to the probability stamps of symbol; Output bit flow is determined according between described code area.

Exemplary, described coding interval range is 2 ⁸~ 2 ⁹.

Image compressing device described in the various embodiments described above is equally for performing the method for compressing image described in the various embodiments described above, and the technique effect of its know-why and generation is similar, is not repeated here.

Note, above are only preferred embodiment of the present invention and institute's application technology principle.Skilled person in the art will appreciate that and the invention is not restricted to specific embodiment described here, various obvious change can be carried out for a person skilled in the art, readjust and substitute and can not protection scope of the present invention be departed from.Therefore, although be described in further detail invention has been by above embodiment, the present invention is not limited only to above embodiment, when not departing from the present invention's design, can also comprise other Equivalent embodiments more, and scope of the present invention is determined by appended right.

Claims (10)

1. a method for compressing image, is characterized in that, comprising:

Preliminary treatment is carried out to original image data;

The view data obtained for preliminary treatment carries out Walsh WALSH conversion;

Entropy code algorithm is adopted to carry out image compression for the view data after conversion.

2. method according to claim 1, is characterized in that, carries out preliminary treatment comprise original image data:

The original image data of RGB rgb format is converted to the view data of yuv format;

Piecemeal process is carried out to the view data of yuv format.

3. method according to claim 1 and 2, is characterized in that, adopts entropy code algorithm to carry out image compression comprise for the view data after conversion:

Quantification treatment is carried out to the view data after conversion;

Zigzag scanning is carried out to the view data after quantizing, obtains corresponding image vector matrix;

Arithmetic Coding algorithm is adopted to encode to image vector matrix.

4. method according to claim 3, is characterized in that, adopts Arithmetic Coding algorithm to carry out coding to image vector and comprises:

The probability that in described image vector matrix, each symbol is corresponding is calculated by probability Estimation;

According to each symbol of described probability stamps;

Determine between code area according to the probability stamps of symbol;

Output bit flow is determined according between described code area.

5. the method according to any one of claim 4, is characterized in that, described coding interval range is 2 ⁸~ 2 ⁹.

6. an image compressing device, is characterized in that, comprising:

Pretreatment module, for carrying out preliminary treatment to original image data;

Conversion module, the view data for obtaining for preliminary treatment carries out Walsh WALSH conversion;

Compressed encoding module, for adopting entropy code algorithm to carry out image compression for the view data after conversion.

7. device according to claim 6, is characterized in that, described pretreatment module is used for:

The original image data of RGB rgb format is converted to the view data of yuv format; Piecemeal process is carried out to the view data of yuv format.

8. the device according to claim 6 or 7, is characterized in that, described compressed encoding module comprises:

Quantification treatment unit, for carrying out quantification treatment to the view data after conversion;

Scanning element, for carrying out zigzag scanning to the view data after quantification, obtains corresponding image vector matrix;

Arithmetic coding unit, encodes to image vector matrix for adopting Arithmetic Coding algorithm.

9. device according to claim 8, is characterized in that, described arithmetic coding unit specifically for:

The probability that in described image vector matrix, each symbol is corresponding is calculated by probability Estimation; According to each symbol of described probability stamps; Determine between code area according to the probability stamps of symbol; Output bit flow is determined according between described code area.

10. the device according to any one of claim 9, is characterized in that, described coding interval range is 2 ⁸~ 2 ⁹.