CN100544398C - The method and system of the adaptability compression of multi-level images - Google Patents
The method and system of the adaptability compression of multi-level images Download PDFInfo
- Publication number
- CN100544398C CN100544398C CNB2005100700194A CN200510070019A CN100544398C CN 100544398 C CN100544398 C CN 100544398C CN B2005100700194 A CNB2005100700194 A CN B2005100700194A CN 200510070019 A CN200510070019 A CN 200510070019A CN 100544398 C CN100544398 C CN 100544398C
- Authority
- CN
- China
- Prior art keywords
- image
- color
- pixel
- information
- prospect
- 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.)
- Active
Links
Images
Landscapes
- Compression Of Band Width Or Redundancy In Fax (AREA)
Abstract
The present invention is convenient to compress such as the adaptability through catching digital picture or the like multi-level images of blank, and this multiple-level encoding comprises the bit stream of color component image and black and white image assembly.Based on user's needs, receiving system ability or the like, any of coloured image and black and white image or both can be the output to the user.
Description
Quoting of related application
The application statement number is 60/567,150 to be entitled as " Adaptive CompressionofMulti-level Images " and the priority of the United States Patent (USP) submitted on April 30th, 2004 to U.S. Provisional Patent Application, and this application is this complete being incorporated herein by reference.
Technical field
The present invention relates to the communication in the computer environment, relate in particular to the compression of the data that are associated with multi-level images.
Background technology
Because to evolution low-cost, high performance communication, solve a problem provides a kind of to one's profit, time-saving method to alleviate execution such as communicating by letter, pay the bill, buy, make a budget and the pressure of daily task such as acquisition of information with entertainment systems to computer and network technologies from the data handling system of expensive, low performance.For example, receive the computing system of internet, can provide almost visit immediately to the user at user's finger tip from the abundant information of the thesaurus of website that spreads all over the world and server by wired or wireless technical battery.
The seizure of image and propagation are gradually dense fields of interest in the computer industry.Three types image or bitmap are arranged usually.For example, full color images comprises on each pixel millions of kinds possible color.Each pixel have three-dimensional hue coordinate R, G, B}, each coordinate is with the 8-16 byte representation.Usually show the full color bitmap with JPEG (associating still image expert group), JPEG2000 or PNG (portable network image technology) form.The color map image comprises some possible on each pixel colors.Usually each pixel has and comprises every kind of color { R, G, the index value in the color form of B} value.The color map image is usually represented with GIF (BIIF) or PNG form.Secondary image (fax) adopts two kinds of colors of every pixel (normally black and white) and is used in file and picture usually.Like this, often show the secondary image with TIFF (TIF) or PNG form.
With the system and method that strengthens the image demonstration demand of failing to satisfy is arranged to being convenient to improve image compression in this area.
Summary of the invention
For the basic comprehension to some aspect of the present invention is provided, below provide simplified summary of the present invention.This " summary of the invention " part is not to be extensive overview of the present invention.It is not to be intended to identify key of the present invention or important element or cropping scope of the present invention.Its sole purpose is to introduce some notion of the present invention with reduced form, as the following prelude that is described in more detail that provides.
The present invention of this announcement and statement therein on the one hand in, relate to the system of being convenient to by the compression of blank and/or document image that capture system produces.Usually, three types image is arranged, each all carries out optimal compression by coder-decoder (codec) or algorithm at such image.The invention provides a kind of color map image that is used for especially by the coder-decoder of catching the color map image that whiteboard images produced.Usually, for such image, all have powerful connections look (for example white) and fraction pixel of most of pixels has the color of can be approximate by one of some colors (for example can choose) and coming from the pre-color form.The color form can have less relatively tone, just is limited to several colors, tone of manufacturing or the like as typical dry-erase markers.Whiteboard images is considered to be multi-level images, and wherein each pixel value is the index (for example " level ") to less color-values form.
According to an aspect of the present invention, provide the compression algorithm that is used for the effective expression of multi-level images by multilevel coding decoding (MLC).Usually each pixel of color map image may have some colors.Each pixel often has and comprises every kind of color { R, G, the index value in the color form of B} value.General color map image encoding decoder comprises GIF and PNG.MLC is convenient to the compression than general GIF or the better color map image of PNG form.In addition, typical encoded bit stream comprises all components that can decode respectively, represents the black and white and the colored version of this image respectively.Thereby when limiting bandwidth, can be from the decode black and white version of this image of the sub-fraction of encoded bit stream.
According to related fields of the present invention, can provide two compression functions to make the compression of input picture to carry out in two steps.At first, to the background/foreground mark of each pixel compress with the black and white of effective compressed image represent-this can finish by for example secondary coding and decoding compressor reducer.Then, can compress by entropy coder (for example from Ziv-Lempel (LZ) series) the color index information of prospect pixel.Of the present invention this is convenient to the enhancing that for example teleconference cooperation, information exchange, blank content are shared or the like on the one hand.In addition, MLC promotes the quality image compression to alleviate possible harmful bandwidth constraints.If bandwidth is subjected to great restriction, the present invention can split into black and white and colored version to encoded information, only transmits the black and white version then to defer to bandwidth constraints (for example through hardwire, wireless or the like transmission).
According to a further aspect in the invention, (m n) encodes to foreground/background image b for available secondary encoder such as secondary coder-decoder (BLC).The size of optionally reducing color transformed table by vector quantization or similar procedure.Encode by the available entropy coder of character string (or vector) that all prospect pixel values form such as the LZ encoder.For example, (m n), and utilizes based on the zlib storehouse of the LZ77 forecolor of encoding can to adopt the BLC encoder to come coding monochrome image b.In conjunction with the present invention, also can utilize the secondary encoder (for example TIFF, PNG) and the string encoding device (for example other LZ mutation) of other kind.
In order to realize aforementioned content and related objective, some illustrative aspect of the present invention is described in detail in conjunction with following specification and accompanying drawing.Yet these aspects only show the wherein several of variety of way that all principles of the present invention can adopt therein, and the present invention is intended to comprise all these aspects and equivalents thereof.Read following detailed description the in detail in conjunction with the accompanying drawings, other advantage of the present invention and new feature will become clear.
Description of drawings
Fig. 1 illustrates the system of being convenient to according to an aspect of the present invention the adaptability compression of the multi-level images such as image of catching such as the digitlization of blank 100.
Fig. 2 illustrates is convenient to the system 200 that the adaptability to the compressed multi-level images such as image of catching such as the digitlization of blank decompresses according to an aspect of the present invention.
Fig. 3 illustrates the system of being convenient to according to an aspect of the present invention the adaptability compression of the multi-level images of the image of catching such as the digitlization of blank etc. 300, and it comprises makes artificial intelligence (AI) assembly of inferring.
Fig. 4 illustrates the Typical Digital image that is produced by the blank capture system.
Fig. 5 illustrates the black and white image that is produced corresponding to Fig. 4 coloured image.
Fig. 6 illustrates the method 600 of being convenient to make up from coloured image black and white image according to an aspect of the present invention by the iterative feedback circulation of each pixel in the estimation coloured image.
Fig. 7 illustrates according to an aspect of the present invention was convenient to the method 700 the adaptability of multi-level images compressed with the reduction of optional color table before compression.
Fig. 8 illustrates the method for being convenient to according to an aspect of the present invention the adaptability compression of the multi-level images such as image of catching such as the digitlization of blank 800.
Fig. 9 and 10 illustrates the example calculation environment of the aspects according to the present invention.
Embodiment
The present invention describes with reference to accompanying drawing, wherein runs through the similar mark mark of institute's drawings attached and represents like.For laying down a definition, statement has a lot of details so that complete understanding of the present invention is provided in the following description.Yet it is obvious not having these details also can put into practice the present invention.In other example, for ease of describing the present invention, well-known structure and equipment illustrate with the block diagram form.
As used in this application, term " assembly " and " system " are intended to refer to computer related entity: the combination of hardware, hardware and software, software or executory software.For example, assembly can be, but is not limited to, and runs on thread, program and/or the computer of process on the processor, processor, object, executable program, execution.As an illustration, application program and the server that runs on the server all can be used as assembly.One or more assemblies can reside in the thread of process and/or execution, and assembly can be positioned on the computer and/or is distributed on two or many computers.
Term " deduction " generally is meant from the process by reasoning or inference system, environment and/or state of user one group of observation of incident and/or data capture when this uses.Infer to can be used to discern specific environment or action, or for example produce probability distribution state.This deduction is a chance-promptly, be concerned about that the probability distribution of state calculates the consideration that is based on data and incident.Infer the technology that also can refer to be used for from one group of incident and/or data, forming the higher level incident.This deduction has caused based on one group of observed event and/or storage event data the structure of new events or action, no matter whether closely property is relevant around for these incidents, and whether these incidents and data from one or several incidents and data source.
The invention provides a kind of coder-decoder, it is used for the color map image especially by catching the color map image that produces, for example coloured image of blank (referring to Fig. 4).These images comprise have background colour most of pixel of (as white) and the fraction pixel with the color of can be approximate by one of some colors in the less color table and coming.This color table can have less relatively tone, just is limited to several colors, tone of manufacturing or the like as typical dry-erase markers.Whiteboard images is considered to be multi-level images, and wherein each pixel value is the index (for example " level ") to less color-values form.
Usually each pixel of color map image may have some possible colors.Each pixel often has { R, G, the index value of the color form of B} value that comprises every kind of color.The coder-decoder of general color map image comprises GIF and PNG.Multilevel coding decoding (MLC) promotes the compression than general GIF or the better color map image of PNG form.In addition, typical encoded bit stream comprises all components that can decode respectively, represents the black and white and the colored version of this image respectively.Thereby when limiting bandwidth, can be from the decode black and white version (referring to Fig. 5) of this image of the sub-fraction of encoded bit stream.
Fig. 1 illustrates the system of being convenient to according to an aspect of the present invention the adaptability compression of the multi-level images such as image of catching such as the digitlization of blank 100.Control assembly 102 can be coupled with compression assembly 104 in operation, and this compression assembly 104 provides the multi-level images of two compression functions with the blank of compression such as numeral seizure to system 100.Compression assembly 104 comprises the encoder 106 that adopts multilevel coding decoding (MLC), and this MLC is the algorithm that is used to compress multi-level images.Encoder component 106 produces the compressed expression (bit stream and/or file) of original image usually, it is made up of several portions: heading message (information that comprises relevant graphical rule, number of colours and other similar parameters), encoded background and encoded foreground character string, and as described below.
For example, can to the 100 input picture x of system (m, n), 0≤m≤N
R-1,0≤n≤N
C-1 (N wherein
RAnd N
CBe respectively line number and columns).(m n) can have the interior value v of scope [0...L-1] to each pixel x, and wherein L is the quantity of color.Real color-values all is stored in color table M={[r
0g
0b
0], [r
1g
1b
1] ..., [r
L-1g
L-1b
L-1], wherein the RGB tristimulus values of determining its corresponding color are all arranged for each color index 0≤v≤L-1.Thereby the input data are that (m is n) with color table M for pixel value x.Compression assembly 104 and control assembly 102 can be determined the backcolor v of input picture
BBecause most of pixels are background colours, can be with v
BBe set at and equal the color-values v that presented by the maximum pixel.Common v
BApproach white, but other background colour also is possible.Can be and include N
R, N
CAnd N
BThe compressed bit stream (or file) of value is write a title.The space of the color table that this title will define after can being designed to reserve.
B (m, n)=0, if x (m, n)=V
B(for example, pixel is on background)
B (m, n)=1, if x (m, n) ≠ V
B(for example, pixel is on prospect)
Available then secondary coder-decoder (BLC) is encoded to black and white image, and can append to the bit that is produced by BLC in output bit flow and/or the file.(m, example n) as shown in Figure 5 corresponding to the black and white image b of image among Fig. 4.In addition, the color-values of encoder 106 codified prospect pixels.Because (m n) learns the position of prospect pixel, can append to character string S={v to the value of all prospect pixels by scanning sequency from black and white image b
0, v
1..., v
F-1In, wherein F is that (usually F is than the total amount N of pixel for the quantity of prospect pixel
RN
CMuch smaller).
For coded string S, systematic encoder 106 can adopt two kinds of patterns: " can't harm " pattern, and wherein use the entropy coder of character string that is suitable for comprising repeat pattern to come the prospect value among the S is carried out direct coding such as Ziv-Lempel (LZ) encoder; And " diminishing " pattern, wherein color table M is reduced to the color table M ' with less color.Then before entropy coding, can (m, n) primary index in replaces to the index of new color table with image x.If required mode can't harm, then character string S is encoded and the bit that obtains is additional to output bit flow or file with the LZ encoder.Color table M can be written in the title of output bit flow or file to finish coding then.
If required mode diminishes, do not need can to suppose the perfect reconstruction of the color of prospect pixel, thereby color just can be similar to, as long as the advancing colour colour display screen must be enough similar to original color when showing.For example, in diminishing pattern, there are input parameter K, required number of colours L, wherein K＜L to encoder 106.Then system 100 can produce new color table M '=[r '
0G '
0B '
0], [r '
1G '
1B '
1] ..., [r '
K-1G '
K-1B '
K-1], thereby for every kind of color among the original table M [r g b],, make new color [r ' g ' b '] enough similar to original color [r g b] in reduction table M ', having color [r ' g ' b '].The reduction of this color table is an example of vector quantization problem: Vector Groups is replaced with than group, make that are good approximation than group's discal patch order in the original set those.Thereby in order to produce new table M ', compression assembly 104 and/or encoder 106 can adopt the vector quantizer algorithm for design such as LBG (Linde-Buzo-Gray is also referred to as the K-intermediate value in statistical theory) or octa-tree tree algorithm.
In case systematic encoder 106 has produced table M ', can adopt the full search vector quantization process of standard that color is redistributed to index.For each prospect pixel, original value v can be replaced by new value v ', for example passes through:
For b (m, n)=1 all (m, n):
With x (m, n)=v ' replace x (m, n)=v, make color [r ' g ' b '] corresponding to v ' be in all colors of reduction color table M ' with original [r g b] color (corresponding to v) immediate color.
Form the character string S that will encode, wherein S={v ' by encoder 106 then
0, v '
1..., v '
F-1And wherein F be the quantity of prospect pixel.Because reduced the scope of the value of each element among the S, will be compressed to the bit stream shorter corresponding to lossless mode than original S.Thereby in diminishing pattern, the bigger reduction of compressed document size is convenient in the less reduction in the reconstruction fidelity.At this moment, can write the title of output bit flow to color table M ' to finish coding.
Fig. 2 illustrates the system of being convenient to according to an aspect of the present invention the adaptability decompression of compressed multi-level images 200.Depend on applicable cases, provide encoder system 100 or decoder system 200 or both.For example, printer driver or compressed document generator can only comprise encoder system 100; File checks that utility program or printer control software can only comprise decoder system 200, and the plug-in card program in document/image processing/tag application program can comprise encoder system 100 and decoder system 200 both.
Now referring to Fig. 3, system 300 illustrates the adaptability compression of being convenient to multi-level images according to an aspect of the present invention.System 300 comprises the control assembly 302 of being convenient to carry out the information analysis that received by system 300.Control assembly is associated in operation with compression assembly 304, and compression assembly comprises encoder 306.
In addition, control assembly 302 comprises the AI assembly 308 that can do about the deduction of system's 300 operations.For example, the AI assembly can be inferred the suitable pattern (for example harmless or diminish) of coding according to relevant user's needs, the information that receives the display capabilities or the like of user's set.According to another example, suitable color form of pixel color or the like in the AI assembly 308 deducibility input modules.In addition, can adopt AI assembly 308 to do judgement and/or deduction about for example bandwidth, if make that bandwidth is subjected to bigger restriction, the information that then need not relevant prospect pixel color just can transmit the black and white version of input picture.Like this, system 300 can provide the meaningful expression of original image and evade bandwidth constraints to receiving the user.
According to related fields, the type of device that receives compressed image (for example being based in part on the information that is associated with request from receiving system) can be judged and/or infer to AI assembly 308, and can adopt such deduction to judge whether this device can receive and/or show multistage coloured image.Can not receive and/or show multi-level images (for example because temporary transient unavailability of bandwidth constraints, device capability restriction, receiving system or the like) if determine receiving system, then system 300 can send the secondary image automatically so that the restriction of compensation receiving system and still provide the information with meaning to receiving the user.
Can comprise for example deduction of relevant user preferences by other example that the work that AI assembly 308 provides is inferred.According to an example, but the multistage version of full color of original image, the multistage version through reducing color and/or secondary, the black and white version of input picture are wished and/or can be received to the hobby of AI assembly 308 assesses user input, information or the like whether so that infer the user.For example, the user who receives compressed image by mobile phone in the remote zone can be set to " roaming " by mobile phone, and AI assembly 308 can be discerned the limited receiving ability of roaming mobile phone, to allow to make the relevant deduction that will send the compressed image type of mobile phone to.In addition, can provide by the relevant information of the receiving ability of 308 apparatus for predicting of AI assembly, make and to point out the user to change the setting in the receiving system so that allow the higher or lower image of the quality of reception to the user.
In addition, system 300 can be shown to the user to the user interface that side information is provided, these information comprise such as the document size that is associated with the compressed image of different quality level, about the user-selectable options that sends and/or receive resource, send priority, file content, as receiving system can force function download time, reduce download time and/or increase suggestion of receiving ability or the like.Check that for strengthening the user experiencing necessary any and all information all offers the user by user interface.Being appreciated that precedent is actually illustrative, is not to be intended to limit the scope that AI assembly 308 is made the mode of deduction, perhaps the deduction quantity that can be done by AI assembly 308.
Term " deduction " generally is meant from the process by reasoning or inference system, environment and/or state of user a series of observations of incident and/or data capture when this uses.Infer to can be used to discern specific environment or action, or produce for example probability distribution of each state.This deduction is that chance-promptly, the probability distribution of correlation behavior is calculated the consideration that is based on data and incident.Infer the technology that also can refer to be used for from a series of incidents and/or data, forming the higher level incident.This deduction has caused based on a series of observed events and/or storage event data structure new events or action, and no matter closely whether property is relevant around for these incidents, and whether these incidents and data from one or several incidents and data source.Carrying out in conjunction with the present invention automatically and/or aspect the action of inferring, can adopt various classification schemes and/or system (for example, support vector machine, neural net, expert system, bayesian belief networks, fuzzy logic, data fusion engines ...).
In order to illustrate Figure 4 and 5 are shown, wherein the illustration numeral is caught the typical whiteboard images (Fig. 4) that can present afterwards.Colored lines among Fig. 4 are represented by different grey lines.When being subjected to bandwidth constraints, can represent black and white image (Fig. 5) corresponding to coloured image (Fig. 4) although so that be that finite bandwidth and/or limited receiving system ability still can be carried out information exchange with the user to the user.For example, in having caught whiteboard images and content thereof and having sent it to incident such as the receiving system of the minimum graph ability of the tool of the handheld apparatus that has simple lcd screen, facsimile machine or the like etc., but the black and white image that initialization system provides handheld apparatus to understand easily and to present.
Although for ease of simplicity of explanation, for example be shown and be described as a series of actions at this with the one or more methods shown in the flow chart form, be to be understood that and think the order that the present invention is not limited to move, can take place or take place synchronously by different order with other action in this description and demonstration because partly move according to the present invention.For example, those skilled in the art will understand and think that method can be expressed as in addition such as a series of inter-related state or incident in state diagram.In addition, be that the not all action that is illustrated of implementation method all is necessary according to the present invention.
Fig. 6 shows and is used for according to an aspect of the present invention catching the method 600 of the adaptability compression of image or the like multi-level images such as blank through numeral.602, can read input picture x (m, n) and the color table that is associated.(m n) and behind the color table that is associated, can begin background process, and it is included in each action that illustrates on the 604-608 to have read input picture x 602.604, determine about the highest color vB=background colour of the frequency of occurrences.606, can determine black and white image b (m, n), 608 it can pass through the secondary encoder encodes.Background process is described in more detail as shown in Figure 7.Be appreciated that if needed only be black and white image then the perspective process that need not the following stated just can realize the method shown in 602-608, thereby can be based on (m, the data that n) are associated are write out output bit flow by the image b with the secondary encoder encodes.
Method can proceed to 610 then, and wherein the prospect pixel value can add among the character string S based on M '.612, the character string S of prospect colour can encode by entropy coder.At last, 614, can be used to write out output bit flow from the encoded character string S of 612 steps with from the encoded monochrome image information of 608 steps.Thereby 610 and 612 be illustrated in the action that takes place during the perspective process, and wherein perspective process is further described with reference to Fig. 8.
Fig. 7 shows the adaptability compression of multi-level images according to an aspect of the present invention and converts the method 700 of secondary image to.702, determine in the relevant image color v of frequent appearance
BThereby the color vB of frequent appearance is appointed as " background " color.For example, input picture can comprise one group of pixel x (m n), makes 0≤m≤N
R-1,0≤n≤N
C-1 (N wherein
RAnd N
CBe respectively line number and columns).Each pixel all has the color-values v in the scope [0...L-1] in the image, wherein the number of colours that expresses possibility of L.The real color-values that is associated with pixel can be stored among the color table M, wherein M={[r
0g
0b
0], [r
1g
1b
1] ..., [r
L-1g
L-1b
L-1] and for each color index 0≤v≤L-1, the RGB tristimulus values of determining its corresponding color are arranged all.For example, if the image of compression and conversion is through filling out the digital picture of picture blank, then most of pixel of comprising of this image will represent white.If like this, white will be endowed v
BTherefore, can calculate to determine background colour v by for example block diagram
B: calculate the occurrence number of every kind of color, and the color that will have a maximum count value is given v
B
704, begin that (m, n) Zu circulation, and 706 are determined whether to represent about each single pixel and equaled v by all pixel x in the image
BValue.If (m, n) (m n) has and v the fixedly pixel x on to determine ad-hoc location
BThe color-values that equates, (m n) is set to 0 as respective background mark value b.If determine pixel x (m, n) color-values and v
BUnequal, then (m n) is set to 1, and promptly value " 1 " is represented non-background colour at the respective background mark value b of 710 particular pixels.In other words, (m represents that n)=1 (m is n) corresponding to the prospect pixel in the position to b.As needs, can comprise value N to compressed bit stream and/or file edit
R, N
CAnd v
BTitle.In addition, the available appointment color epiphase additional space of closing information makes up title.
In case particular pixels estimated and given value, then can 712 judge whether to handle pixels all in the image (m, n).If determine not estimate as yet that 712 all pixels (then can get back to 704 and make another time pixels estimation iteration and (promptly have new pixel addresses to (m, n)) for m, collection n) by method.Yet, if 712 determine to have handled all pixels (m, n) collection, can with original image x (m, n) be mapped to corresponding black and white image b (m, n), if wherein x (m, n)=v
BThen b (m, n)=0, and if x (m, n) ≠ v
BThen b (m, n)=1.Method proceeds to 714 then, can come wherein that (m n) encodes to black and white image b by for example secondary encoder.The bit of secondary encoder encodes can be appended in output bit flow and/or the file then.
Fig. 8 illustrates and is used for according to an aspect of the present invention catch the method 800 of the adaptability compression of image or the like multi-level images such as the digitlization of blank.802, determine about during with the data that are associated through the seizure image, whether needing to diminish pattern in compression.For example, typically diminish the data compression coding decoder in conjunction with flow medium, phone or the like employing.According to an example, can adopt to diminish the transform coding decoder, one of them image pattern can be selected, be split into smaller portions, convert new fundamental space to and quantize at last and entropy coding.According to another example, can adopt in conjunction with the present invention to diminish the predictive coding decoder, can use the coder-decoder of early decoding to predict present image by it.Can quantize and be coded in the error between prediction and the real data then.According to another example, can adopt the combination of conversion and prediction lossy coding decoder.
If selected to diminish pattern, then method proceeds to 804, and wherein color table M can reduce by vector quantization technology, octa-tree reduction, centre of moment analytical technology or the like.The pattern that need diminish is improved the compression of complex figure.For example, can use the single color (for example blue) of planting on the white background to fill out the picture blank.If as once being rewritten as on other mark (as redness, black or the like) with this blue markings pen of usually taking place, the blue markings that then is used for filling out the picture blank can be deposited on the blank on these other colors.Can leave out to have and represent the required pure minimum pixel of true blue color marker on the white background to provide with two kinds of different color-values pixels of color-values (for example white background and blue markings) that occur the most frequent.Be appreciated that the specific color that provides do not make limited significance and explain in precedent because except blank among the present invention other backcolor and/or mark, other image subject also be possible.
In addition, if blank is filled out picture (some is main) with multicolour, can adopt the pattern of diminishing that color table M is reduced to M ', wherein M ' comprises the collection of the main color of great majority.For example, if main blue, green and the red-label pen of using filled out the picture blank, and use the orange marker pen to come the picture single line, the color table M that then comprises blueness, green, redness and orange can be reduced to and comprise blueness, redness and green M '.According to this situation, the orange lines can be with the representing of redness, because redness is at the most close pixel color in reduction color table M '.
806, prospect pixel character string can be initialized as " sky.In addition, if determine not need to diminish pattern 802, then method can directly proceed to 806, is used for not reducing color table and initialization prospect pixel character string S.Then, 808, can to all pixels in the image (m, n) begin the circulation.810, determine to determine whether (b (m, n)=1) for example in prospect of given pixels about reference method 700.If determining this pixel is not that (for example, it is background pixels b (m, n)=0) to the prospect pixel, and then method is got back to 808 and is used for remaining pixel is made further iteration.If determine that 810 pixel is the prospect pixel, then determine whether the pattern that to diminish 812.Diminish pattern if desired, then 814 with pixel (m, color-values c n) replace to the reduction color table M ' that hangs oneself near color c.816, (m, new color-values c n) appends among the character string S pixel.
Not need in the incident of the pattern that diminishes to determine 812, method directly proceeds to 816 from 812, and wherein (m, color-values c n) will be affixed among the character string S pixel.Then, 818, can determine whether to have handled all pixels (m, n).If only handled the part among the pixel complete or collected works, then method turns back to 808 to make other picture element analysis iteration.If determine to have estimated all pixels that then method can proceed to 820 818, wherein the character string S of forecolor value can encode by entropy coder (for example from Ziv-Lempel (LZ) series).
Should be noted that Fig. 6,7 and 8 illustrates the flow chart 600,700 and 800 of the preferred embodiment of the adaptability compression that is used for multi-level images according to the present invention.For those skilled in the art, it is possible using the mutation of flow chart 600,700 and 800, thereby operating sequence also can be changed into and for example is more suitable for particular hardware platform; These mutation remain embodiments of the invention.
It should be noted that flow chart 600,700 and 800 specified the preferred embodiment of encoder 406.From which flow chart in the operation of appointment, can the derive respective flow chart of encoder 206 of those skilled in the art.
Referring now to Fig. 9, show the block diagram that to operate the computer of carrying out framework of the present invention.For other context of aspects of the present invention is provided, Fig. 9 and following argumentation be intended to the suitable computing environment 900 that aspects of the present invention is realized therein provide briefly, bright in general.Although the present invention illustrates in the general context of the computer executable instructions on running on one or more computer, those skilled in the art will recognize that the present invention also can realize in conjunction with other program module and/or as the combination of hardware and software.Usually, program module comprise the routine carrying out specific tasks or realize concrete abstract data structure, program, assembly, data structure, or the like.In addition, it will be appreciated by those skilled in the art that method of the present invention also can put into practice by other computer system configurations, comprise uniprocessor or multiprocessor computer system, minicom, mainframe computer and personal computer, hand-held computing device, based on microprocessor or programmable consumer appliances or the like, wherein each device in operation with one or more associated devices couplings.The illustrated aspects of the present invention also can be put into practice in distributed computing environment (DCE), and wherein task is carried out by the teleprocessing equipment that connects through communication network.In distributed computing environment (DCE), program module can place local and remote memory device.
Refer again to Fig. 9, realize that the exemplary environments 900 of each side of the present invention comprises computer 902, this computer 902 has processing unit 904, system storage 906, reaches system bus 908.System bus 908 coupled system assemblies include but not limited to system storage 906 is coupled to processing unit 904.Processing unit 904 can be any of various available processors.Dual micro processor and other multiple processor structure also can be used as processing unit 904.
Numerous program modules comprise operating system 930, one or more application program 930, other program module 934 and routine data 936, can be stored among driver and the RAM 912.Be appreciated that the present invention can realize with the combination of various operating systems of buying or operating system.
The user can by as keyboard 938 and such as the positioner of mouse 940 to computer 902 input instructions and information.Other input unit (not shown) can comprise microphone, IR remote controller, joystick, game mat, satellite receiver, scanner or the like.These and other input unit is connected to processing unit 904 by the input unit interface 942 with system bus 908 couplings usually, but also may link to each other, as parallel port, game port, USB (USB) port, IR interface or the like by other interface.Monitor 944 or other type display device also by interface, link to each other with system bus 908 as video adapter 946.Except display 944, personal computer generally includes other peripheral output devices (not shown), as loud speaker and printer etc.
When being used for the lan network environment, computer 902 is connected with local area network (LAN) 952 by wired and/or wireless communications network interface or adapter 956.Adapter 956 can help the wired or wireless communication with LAN952, comprising the WAP (wireless access point) that is used for communicating by letter with wireless adapter 956.When being used for the WAN network environment, computer 902 generally includes modulator-demodulator 958 or is connected in the communication server on the LAN or has the device that other is used for setting up as the internet at wide area network 954 communication.Can be built-in or the modulator-demodulator 958 of external, wired or wireless device is connected by serial port interface 942 with system bus 908.In networked environment, program module or its part relevant with described computer 902 can be stored in remote memory/storage device 950.It is exemplary that network shown in being appreciated that connects, and also can use other to be used for setting up at intercomputer the device of communication link.
It is exercisable placing the wireless device of radio communication or communicating by letter of entity in computer 902 and any operation, for example printer, scanner, desk-top and/or portable computer, portable digital-assistant, any equipment or place (for example telephone booth, newsstand, lavatory) and phone that is associated with wireless detectable label.This comprises Wi-Fi and bluetooth at least
TMWireless technology.Thereby communication can be the predetermined structure that has special communications between general networks or at least two devices.
Wi-Fi or Wireless Fidelity, make in sofa in the family, the accommodation bed or work in meeting room need not wiring and just can be connected with the internet.Wi-Fi impels the wireless technology of this device of computer for example in this class of picture cellular telephone of indoor and outdoors transceive data; Communication Anywhere that can be in the scope of base station.The Wi-Fi network use the radiotechnics be called IEEE 802.11 (a, b, g etc.) provide safety, reliably, wireless connections fast.The Wi-Fi network can be used for interconnecting between the computer, with being connected of internet, cable network (using IEEE802.3 or Ethernet).The Wi-Fi network has 11 mbit/(Mbps) (802.11a) or the digit rate of 54Mbps (802.11b), or the product that comprises two frequency bands (double frequency-band) operates on the radio band of 2.4 and 5GHz of no licence, so network can provide the actual performance that is similar to the basic 10BaseT wired ethernet that uses in many offices.
Referring now to Figure 10,, shows the schematic block diagram of example calculation environment 1000 according to the present invention.System 1000 comprises one or more client computer 1002.Client computer 1002 can be hardware and/or software (for example thread, process, calculation element).Client computer 1002 as can be by the contextual information that adopts the present invention to contain the cookie program and/or be associated.System 1000 also can comprise one or more servers 1004.Server 1004 also can be hardware and/or software (for example thread, process, calculation element).Server 1004 can adopt the present invention to contain thread and carry out conversion.May communicate by letter at one of 1004 of client computer 1002 and servers may be that data packet form to be suitable for transmitting between two or more computer processes is carried out.Packet can comprise cookie program for example and/or the contextual information that is associated.System 1000 comprises and can be used to be convenient to 1004 communications frameworks of communicating by letter 1006 of client computer 1002 and the server global communications network of internet (for example such as).
Communication can promote by wired (comprising optical fiber) and/or wireless technology.Client computer 1002 can be connected in operation with one or more client data storages 1008 that are used for storage client 1002 local informations (for example cookie program and/or associated context information).Similarly, server 1004 can be with one or more to be used for storing to server 1004 be that the server data stores 1010 of local information is connected in operation.
The above comprises many examples of the present invention.Certainly, be impossible for describing the present invention to each thinkable line description that is combined into of assembly or method, but those of ordinary skills understand that more combinations of the present invention and change are possible.Therefore, the present invention is intended to comprise all and such drop on change, the modification in the appended claims spirit and scope and change.In addition, with regard to the term that is used for embodiment or claims " had ", this term was intended to " comprise " that to be similar to term in claims the mode that connects speech makes to comprise meaning and separate.
Claims (37)
1. a method of being convenient to colored reflection image compression is characterized in that, comprising:
The background/foreground mark of the pixel of compressed color map picture; And
Compress the colored index information of the prospect pixel of described color mapped image; And
Generation is corresponding to the title of the compressed file that is associated with described image, and wherein said title comprises the value v of the background colour of representing described image
B, v
BEqual to represent the color-values that the maximum quantity pixel of described image is presented.
2. the method for claim 1 is characterized in that, comprises that also the black and white that adopts secondary coding and decoding compressor reducer to compress described image represents.
3. the method for claim 1 is characterized in that, also comprises the colored index information that adopts the LZ entropy coder to compress the prospect pixel of described image.
4. the method for claim 1 is characterized in that, also comprises adopting colored index information and/or the background/foreground label information of a decoder optionally to decode and to be associated with the pixel of color mapped image, so that output image is provided.
5. the method for claim 1 is characterized in that, comprises that also the value by corresponding pixel is set to equal or be not equal to v
BFunction, come described image mapped is become corresponding black and white image.
6. method as claimed in claim 5 is characterized in that, also comprises at least in part based on the encode color-values of prospect pixel of the prospect pixel location of determining from described black and white image.
7. method as claimed in claim 6 is characterized in that, the coding of described prospect pixel color-values comprises the value of described prospect pixel is appended among the character string S by scanning sequency.
8. method as claimed in claim 7 is characterized in that, adopts the lossless compression scheme described character string S that encodes.
9. method as claimed in claim 7 is characterized in that, adopts the lossy compression method scheme described character string S that encodes.
10. the method for claim 1 is characterized in that, the compressed color index information comprises carries out the octa-tree reduction.
11. the method for claim 1 is characterized in that, the compressed color index information comprises that execute vector quantizes.
12. method as claimed in claim 11 is characterized in that, execute vector quantizes to comprise employing Linde-Buzo-Gray algorithm.
13. the method for claim 1 is characterized in that, the compressed color index information comprises a kind of technology that is used to make the mean square error minimum between original color and the encoded color.
14. a system of being convenient to the compression of multi-level images adaptability is characterized in that, comprising:
One control assembly, the suitable pattern that it receives digital picture and determines the described image of coding;
One compression assembly, the background/foreground mark that its compression is associated with pixel in the described digital picture; And
One secondary coder-decoder, it is encoded to comprising the background that is associated with described image and the character string of prospect pixel information, wherein the encoded character string of background and prospect pixel information has a title, described title comprise with picture size, background colour, described image in number of colours and the relevant information of color table of described image.
15. system as claimed in claim 14 is characterized in that, described compression assembly also compresses the colored index information that is associated with prospect pixel in the multi-level images.
16. system as claimed in claim 15 is characterized in that, described compression assembly comprises the LZ entropy coder that compresses described colored index information.
17. system as claimed in claim 16 is characterized in that, also comprises making the judgement relevant with system operation and/or the AI assembly of derivation.
18. system as claimed in claim 17, it is characterized in that, described AI assembly adopts at least one in vectorial machine, neural net, expert system, bayesian belief networks, fuzzy logic, the data fusion engines, to carry out judgement and/or the deduction relevant with system operation.
19. system as claimed in claim 14 is characterized in that, also comprises a decoder, it reads described title and encoded character string, and the decode first bit stream component and the second bit stream component.
20. system as claimed in claim 19 is characterized in that, the described first bit stream component comprises the information relevant with the secondary version of original image.
21. system as claimed in claim 19 is characterized in that, the described second bit stream component comprises the information relevant with the colored version of original image.
22. system as claimed in claim 21 is characterized in that, described decoder is given each of a plurality of prospect index color according to described color table.
23. a system of being convenient to the two compressions of multi-level images is characterized in that, comprising:
One control assembly, it analyzes the information that is associated with described multi-level images;
One secondary coding and decoding compressor reducer, its compressed background and prospect pixel are represented with the black and white that makes up described multi-level images; And
One entropy coder, the colored index information that its compression is associated with described prospect pixel, described colored index information has a title, described title comprise with picture size, background colour, described image in number of colours and the relevant information of color table of described image.
24. system as claimed in claim 23 is characterized in that, described entropy coder is the Ziv-Lempel encoder.
25. system as claimed in claim 23 is characterized in that, described control assembly is analyzed bandwidth to determine whether successfully to transmit the black and white or the colored expression of described multi-level images.
26. system as claimed in claim 25 is characterized in that, described control assembly splits into encoded information the black and white and the colored version of described multi-level images.
27. system as claimed in claim 26 is characterized in that, when bandwidth was under predetermined threshold, described control assembly only transmitted the black and white version of described multi-level images.
28. a system of being convenient to the bilingual compression of multi-level images is characterized in that, comprising:
One control assembly, it analyzes the information that is associated with compressed multi-level images; And
One decoder, it reads the background with a title and the encoded character string of prospect pixel information, described title comprise with picture size, background colour, described image in number of colours and the relevant information of color table of described image.
29. system as claimed in claim 28 is characterized in that, described decoder is based on decode from the information of described encoded character string and title the collection first bit stream component and the second bit stream component.
30. system as claimed in claim 29 is characterized in that, the described first bit stream component comprises the information that is associated with the secondary version of original image.
31. system as claimed in claim 30 is characterized in that, also comprises the secondary decoder of the described first bit stream component of decoding.
32. system as claimed in claim 29 is characterized in that, described second bit stream comprises and the relevant information of the colored version of original image.
33. system as claimed in claim 32 is characterized in that, also comprises the entropy decoder of the described second bit stream component of decoding.
34. system as claimed in claim 28 is characterized in that, described decoder is given each of a plurality of prospect index color according to described color table.
35. a system of being convenient to the compression of multi-level images adaptability is characterized in that, comprising:
Be used to receive the device of image;
The device of prospect and background pixels of described image is used to encode; And
The device that is used to the value coded string that is associated with described prospect pixel; And
Be used to produce the device corresponding to the title of the compressed file that is associated with described image, wherein said title comprises the value v of the background colour of representing described image
B, v
BEqual to represent the color-values that the maximum quantity pixel of described image is presented.
36. system as claimed in claim 35 is characterized in that, comprises that also being used for a color table M who is associated with described image pixel color reduces the device that becomes the color table M ' that is associated with the color-values of described prospect pixel.
37. system as claimed in claim 36 is characterized in that, the described device that is used to reduce described color table M comprises the one at least of vector quantization technology and octa-tree technology.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US56715004P | 2004-04-30 | 2004-04-30 | |
| US60/567,150 | 2004-04-30 | ||
| US10/916,721 | 2004-08-12 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1694486A CN1694486A (en) | 2005-11-09 |
| CN100544398C true CN100544398C (en) | 2009-09-23 |
Family
ID=35353277
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100700194A Active CN100544398C (en) | 2004-04-30 | 2005-04-29 | The method and system of the adaptability compression of multi-level images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100544398C (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008019524A1 (en) * | 2006-07-17 | 2008-02-21 | Thomson Licensing | Method and apparatus for encoding video color enhancement data, and method and apparatus for decoding video color enhancement data |
| CN108111858B (en) * | 2016-11-24 | 2020-06-05 | 腾讯科技（深圳）有限公司 | Picture compression method and device |
| CN106909970B (en) * | 2017-01-12 | 2020-04-21 | 南京风兴科技有限公司 | Approximate calculation-based binary weight convolution neural network hardware accelerator calculation device |
| CN107480437B (en) * | 2017-08-01 | 2020-09-18 | 西安万像电子科技有限公司 | Data transmission method and device |
-
2005
- 2005-04-29 CN CNB2005100700194A patent/CN100544398C/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN1694486A (en) | 2005-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1598781B1 (en) | Adaptive compression of multi-level images | |
| CN109151468B (en) | Image data encoding and decoding methods and devices | |
| CN106899861B (en) | A kind of photograph document handling method and its equipment, system | |
| US8594137B2 (en) | Apparatus and methods for image decoding | |
| JP5461419B2 (en) | Pixel predicted value generation procedure automatic generation method, image encoding method, image decoding method, apparatus thereof, program thereof, and recording medium on which these programs are recorded | |
| US20140072239A1 (en) | Image Compression Using Sub-Resolution Images | |
| CN105100814B (en) | Image coding and decoding method and device | |
| CN103918261A (en) | Signal processing and inheritance in a tiered signal quality hierarchy | |
| EP1952539A1 (en) | Repetition and correlation coding | |
| CN100544398C (en) | The method and system of the adaptability compression of multi-level images | |
| WO1998003008A1 (en) | Method for transmitting multiresolution image data in a radio frequency communications system | |
| CN111654721A (en) | Video processing method, system, electronic device and storage medium | |
| CN106133791A (en) | Clustering and coding for color compressed | |
| CN102811347A (en) | Image processing device, image processing method, and program | |
| CN115474058A (en) | Point cloud encoding processing method, point cloud decoding processing method and related equipment | |
| CN101919248A (en) | Byte representation for enhanced image compression | |
| CN116325731A (en) | Point cloud encoding and decoding method and system, point cloud encoder and point cloud decoder | |
| US20020001414A1 (en) | System for building a data compression encoder | |
| JP4411411B2 (en) | Data compression apparatus and data compression method | |
| CN114554205B (en) | Image encoding and decoding method and device | |
| CN105745924A (en) | Chroma down-conversion and up-conversion processing | |
| CN109587502A (en) | A kind of method, apparatus of frame data compression, equipment and computer readable storage medium | |
| JP3145908B2 (en) | Data decoding device and data encoding device | |
| JP2016021600A (en) | Image encoder, image processor, image transmission/reception system, image encoding method, and control program | |
| CN108769695B (en) | Frame type conversion method, system and terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: NIVANCS NETWORKS CO., LTD. Free format text: FORMER OWNER: MICROSOFT CORP. Effective date: 20091211 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20091211 Address after: Delaware Patentee after: Niphan Zahn Network Co. Ltd. Address before: Washington State Patentee before: Microsoft Corp. |