CN101635853A - Method for processing a digital image - Google Patents
Method for processing a digital image Download PDFInfo
- Publication number
- CN101635853A CN101635853A CN 200910160228 CN200910160228A CN101635853A CN 101635853 A CN101635853 A CN 101635853A CN 200910160228 CN200910160228 CN 200910160228 CN 200910160228 A CN200910160228 A CN 200910160228A CN 101635853 A CN101635853 A CN 101635853A
- Authority
- CN
- China
- Prior art keywords
- coefficient
- image
- block
- bit stream
- bitstream
- 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
- 238000000034 method Methods 0.000 title claims description 77
- 238000012545 processing Methods 0.000 title description 17
- 230000006835 compression Effects 0.000 claims description 43
- 238000007906 compression Methods 0.000 claims description 43
- 238000003860 storage Methods 0.000 claims description 29
- 238000005267 amalgamation Methods 0.000 claims description 20
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000013139 quantization Methods 0.000 claims description 6
- 230000003068 static effect Effects 0.000 claims description 6
- 238000012856 packing Methods 0.000 claims description 2
- 238000004513 sizing Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 abstract description 10
- 230000015654 memory Effects 0.000 description 19
- 210000003128 head Anatomy 0.000 description 7
- 238000006073 displacement reaction Methods 0.000 description 5
- 238000013459 approach Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 3
- 238000005520 cutting process Methods 0.000 description 3
- 238000003795 desorption Methods 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- 238000011946 reduction process Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005538 encapsulation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 239000012464 large buffer Substances 0.000 description 1
- 230000005039 memory span Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Abstract
An image representation format for representing a digital image comprises: image information stored as a bitstream representing sequential image blocks, each block comprising one or more components, each component comprising one or more data units and each data unit being represented as a Huffman-coded stream of coefficients of basis functions, wherein a zeroth order coefficient is represented as a difference to the previous zeroth order coefficient of the corresponding component, and a block information table, which comprises: indicators to the first coefficient of a specified order of each image block in said bitstream, information indicating the number of bits in the bitstream between adjacent coefficients of said specified order of the image block, and the zeroth order coefficient of at least the first data unit of each component, said zeroth order coefficient being represented in a non-differential form.
Description
The application is that application number is 200480034015.4, the applying date is on November 16th, 2004, denomination of invention is divided an application for the application for a patent for invention of " method of processing digital images and graphical representation form ".
Technical field
The graphical representation form that the present invention relates to the method for processing digital images and be used to represent digital picture.The invention further relates to the method that is used for raw image data is encoded to the digital image representation of compression, and the method for analyzing the digital picture of JPEG compression.
Background technology
In present day society, create the information of enormous quantity every day.Many information present with the form of image.In addition, most information for example stores and is presented on the Internet electronically.Simultaneously, the use of arriving the wireless connections of Internet increases, and wherein the message transmission rate in these wireless connections is low relatively.Therefore, have needs with very compact form presentation information.This is even more important for image, and this is to be represented as sizable data set because be stored as the normal digital image of pel array.
In addition, the mobile phone that the comprises camera increased popularity that just becoming.Therefore, mobile phone is necessary can processing digital images.Mobile phone or other handheld device of handling image have finite storage space and limited disposal ability.Therefore, if on such equipment carries out image processing, then need effectively and intelligently to store digital picture so that memory space and disposal ability are provided with low requirement.
For this reason, compressed image there is very large interest.The popular approach of compressed image is JPEG (joint photographic experts group) standard.In CCITT Rec.T.81, defined Joint Photographic Experts Group.
Yet, for the sake of clarity, provided Short Description below according to the image file format of Joint Photographic Experts Group.
Joint Photographic Experts Group has defined the lossy baseline encoded system based on dct transform, and the extended coding system that is used for presenting with less data volume converted image.When digital picture is converted to the jpeg file form, carry out the dct transform and the quantification of image, wherein each component in the color space model of this image is carried out dct transform respectively.The all colours component is represented as piece, and these pieces are sequentially handled.Piece behind the dct transform is carried out thresholding and quantification, so that abandon the information that the perception of image is not almost had the basic function of influence.Use Huffman (Huffman) coding that the coefficient of zero order (DC coefficient) of each component in each piece is stored as difference with previous DC coefficient.Sequentially arrange the coefficient (AC coefficient) of higher-order, obtain this sequence by zigzag order from array.The AC coefficient is carried out zero run length (run length) coding, and further utilize huffman coding these AC coefficients of encoding.
Developed the jpeg file form so that create a kind of standard compression of storage size of abundant minimizing digital picture.Therefore, the jpeg file form is not suitable for image manipulation.If want processing digital images, then most convenient is this digital service unit to be got back to spatial domain represent.Yet, when on unit such as mobile phone, that have little memory space, handling image, may not necessarily handle the large buffer memory requirement of the digital picture of representing with spatial domain.
EP 1037165 has described a kind of method that is used to operate with the digital picture of jpeg format storage.For the recognition image zone in the bit stream of jpeg image the position and this bit stream of prescan.A plurality of positions of the appointment in these positions are stored in the prescan form so that by access easily, whereby in the time will operating a part of image, and selection area that can this image of access and do not need the whole bit stream of decoding.Yet, still need when keeping low memory requirement, further improve the speed of image processing.
Summary of the invention
The purpose of this invention is to provide can be easily analyzed and/or the compressed image of operation.Further purpose of the present invention provides easily digital picture amalgamation (stitch) is the possibility of compressed image presentation format.
According to a first aspect of the invention, these and other purpose of the present invention realizes by a kind of method of processing digital images.This method comprises: provide digital picture with compressed format, wherein this digital picture is represented as the bit stream of an expression consecutive image piece, each piece comprises one or more components, each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of basic function coefficient, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component; And provide block information table, it comprises: the zeroth order of each image block in the bit stream or the designator of coefficient of first order, between the zeroth order or coefficient of first order of indication in the adjacent data cell of this image block, the information of the bits number in the bit stream, and the coefficient of zero order of at least one data cell of each component, described coefficient of zero order is represented with the form of non-difference.This method also comprises, for each data cell at least one image block: the coefficient of zero order of this data cell of access, and do not carry out Hofmann decoding to the coefficient of this data cell or to the coefficient of the predetermined number of this data cell, by using the information relevant in the block information table to jump to next zeroth order or coefficient of first order in the bit stream with bits number between the coefficient of adjacent data cell in the bit stream, skip remaining coefficient, decode thus through the simplification collection of the coefficient of huffman coding.
In the context of this application, term " image block " should be considered to the space segment of presentation video, the wherein said information that can have from different color model component.Each image block can be represented as the one or more coefficient sets that are used for each color model component.
According to a second aspect of the invention, be used to represent that by a kind of the graphical representation form of digital picture realizes these purposes.This graphical representation form comprises: the image information that is stored as the bit stream of expression consecutive image piece, each piece comprises one or more components, each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of basic function coefficient, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component; And block information table, it comprises: the zeroth order of each image block in the described bit stream or the designator of coefficient of first order; Between the zeroth order or coefficient of first order of indication in the adjacent data cell of this image block, the information of the bits number in the bit stream; And the coefficient of zero order of at least one data cell of each component, described coefficient of zero order is represented with the form of non-difference.
According to a third aspect of the invention we, provide a kind of method that is used for raw image data is encoded to the digital image representation of compression.This method comprises: the image block that reads raw image data middle finger sizing with any order, and for each image block: this image block is transformed to one or more data cells of one or more component, and described conversion is created as the expression of each data cell the coefficient of basic function; Calculate the quantization approximation of described coefficient; At least some quantization parameters are expressed as the coefficient stream of consecutive image piece; Described coefficient stream is carried out huffman coding, wherein coefficient of zero order is expressed as difference with the previous coefficient of zero order of respective component; The described huffman-coded stream of packing coefficient in bit stream; In block information table, store the zeroth order of each image block in the bit stream or the designator of coefficient of first order into; In block information table between the storage indication zeroth order or coefficient of first order in the adjacent data cell of image block, the information of the bits number in the bit stream; And the coefficient of zero order of at least one data cell of each component of storage in block information table, described coefficient of zero order is represented with the form of non-difference.
According to a forth aspect of the invention, a kind of method of digital picture of the JPEG of analysis compression is provided, the digital picture of this JPEG compression is represented as bit stream, wherein said bit stream is represented the continuous images piece, each piece comprises one or more components, each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of basic function coefficient, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component.This method comprises: sequentially progressively by this bit stream, and progressively by this bit stream the time: store the zeroth order of each image block or the designator of coefficient of first order in block information table into; This coefficient of zero order of decoding, and in block information table the coefficient of zero order of at least one data cell of each component of storage, described coefficient of zero order is represented with the form of non-difference; And in block information table between the storage indication zeroth order or coefficient of first order in the adjacent data cell of image block, the information of the bits number in the bit stream.Progressively the non-zeroth coefficients (wherein said non-zeroth coefficients is represented by the bitstream entries sequence) by the data cell in the bit stream comprises: the bit sequence of checking the subsequent bit of predetermined number in this bit stream; Carry out table lookup, with the category (cathegory) of determining at least the first bitstream entries in this bit sequence and zero run length and the bit length of determining first bitstream entries; Skip in the bit stream with the corresponding bits number of determined bit length; The bits number of being skipped is sued for peace, so that collect and the relevant information of bits number between the zeroth order or coefficient of first order of adjacent data cell, in the bit stream; And the number of coefficients of progressively having passed through sued for peace, up to the whole coefficients that progressively passed through this data cell or till running into the block end symbol.
According to a fifth aspect of the invention, provide a kind of method that is used for two digital pictures of amalgamation (stitch).This method comprises: determine the spatial relationship between these two digital pictures; According to the spatial relationship between these two digital pictures, the image block of distribute digital image information in two digital picture index; Form the bit stream of representing the consecutive image piece according to the index that is distributed, wherein each piece comprises one or more components, and each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of basic function coefficient; Position according to image block will be used for the image block information stores of each image block in block information table, described image block information comprises: the zeroth order of each image block in the described bit stream or the designator of coefficient of first order, between the zeroth order or coefficient of first order of indication in the adjacent data cell of image block, the information of the bits number in the bit stream, and the coefficient of zero order of at least one data cell of each component, described coefficient of zero order is represented with the form of non-difference.
Because at least some aspects of the present invention, digital picture represent that with the graphical representation form of the little memory space of needs the while still can easily be handled with the digital picture of the presentation format of compression.This is particularly useful for application such as mobile phone, that have little memory space and reduction process ability.The present invention allows digital picture to store with compressed format, but still handles and operate with real-time mode when digital picture is presented on the screen.Because according to graphical representation form of the present invention be used to handle the method for image, this image can be very apace presents with the ratio of dwindling or with the resolution that has reduced.The storage of the coefficient of zero order of image block is hinting the information calculations coefficient of zero order that does not need to use previous coefficient of zero order.In addition, can the decode non-zeroth coefficients of any desired number.Can skip remaining non-zeroth coefficients in the bit stream fast, this is because block information table provides the bit length information between the coefficient of adjacent data cell, to allow the fast access to next data cell.This is hinting can be with the ratio of dwindling decoded picture apace, and this is because the non-zeroth coefficients in the bit stream that do not need to decode is searched the beginning of next data cell in the image block or the beginning of next image block.Therefore, this graphical representation form allows the fast access to digital picture.
Store the zeroth order in each image block or the designator of coefficient of first order into, the fast access to some part of image is provided, and has not needed to begin the huffman-coded stream of desorption coefficient from the beginning of stream.On the contrary, can use designator direct access image block.In addition, the coefficient of zero order of at least one data cell in each component form with non-difference in block information table is represented.Therefore, the needs of this value have been avoided from the huffman-coded stream of coefficient, calculating.This allows to present and operate the part of this digital picture when digital picture is in the graphical representation form of compression, this is because the part of access randomly and analysis image.
Graphical representation form in according to the present invention, what have alleviated digital image compression is the target of minimal size.Therefore, do not have the optimally size of compressed digital video, but store some further information of the huffman-coded stream of closing coefficient on the contrary discretely, so that allow some part of rapidly retrieving images.Especially, might pass through the only part of Hofmann decoding non-zeroth coefficients, and use to be suitable for the inverse discrete cosine transform of the smaller piece such as 4 * 4, come very apace a part with another ratio decoded picture or image so that calculate the piece of less size.This graphical representation form can easily be converted to jpeg image, and this is because this graphical representation form is very similar to jpeg image format.Therefore, when image does not need further to operate, might carry out the coefficient value of conversion and the deletion designator and the storage of jpeg image.
The huffman-coded stream of coefficient does not need to comprise the coefficient of zero order that is stored in discretely in the block information table.Yet, in any case the huffman-coded stream of coefficient can comprise whole coefficients.If be other picture format with this graphical representation format conversion, then this may be suitable, and this is because this bit stream self may directly use in another kind of picture format.
Term " block information table " should strictly not regarded as a form, store this fact of stored information in a controlled manner but should only be looked at as, wherein the position of stored information in form is relevant with locus in the represented image of this information.Therefore, this block information table can for example be divided into several tabulations or form.In addition, " huffman-coded stream of coefficient " may not mean that whole stream all is huffman coding.This stream can comprise the Huffman code that mixes with the initial data that is used for coefficient.For example, in the JPEG compressed file, the zero run length of coefficient and category are by huffman coding, and the coefficient value in category is described by not compressed position.
By storage indicator in block information table and coefficient, obtain the structure of knowing of store image information.
Image block can for example comprise three color model component, i.e. a luminance component and two chromatic components.Represent digital picture with a luminance component and two chromatic components, hint degree of staining component can be represented and almost without any the perceptible information dropout of eyes with lower resolution.Therefore, be with regard to three chromatic components with digital image representation just, can realize the initial compression of the information in the digital picture.Use such compression, each image block can for example comprise four data unit and be used for luminance component, and comprises a data unit and be used for each chromatic component.In the huffman-coded stream of coefficient, these data cells are sequentially sorted.
According to a sixth aspect of the invention, provide a kind of graphical representation form that is used to represent digital picture.This graphical representation form comprises: the image information that is stored as the bit stream of expression consecutive image piece, each piece comprises one or more components, each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of basic function coefficient, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component; And the bitstream information of storing together in conjunction with bit stream, described bitstream information comprises the information of the bits number of each data cell in the indicating image piece.
This graphical representation form needs relatively little memory space.This graphical representation form is the holding position stream information only, and the graphical representation form that this bitstream information makes it possible to is according to a second aspect of the invention created block information table apace.Therefore, the longer-term storage that is suitable for image according to this graphical representation form of sixth aspect present invention.When access graph as the time, this bitstream information can be used for analyzing apace this bit stream and creates block information table.The information of the bits number of each data cell can be used for these data cells of access apace in the indicating image piece.When the access data unit, can create designator, can collect indication in the zeroth order of adjacent data cell or the information of the bits number between the coefficient of first order, and this coefficient of zero order of can decoding is so that store at least one coefficient of zero order of each component with the form of non-difference.
Can compress this bitstream information.This is hinting the graphical representation form that can more effectively store in the 6th aspect.
Designator in the block information table can be indicated biased the moving of the coefficient from static position to described appointment order.This static position can be for example beginning of bit stream.Therefore, this designator is implemented as the pointer of the specific bit-position in each image block.This is hinting can retrieve each image block apace, can quicken the access to the specific part of the huffman-coded stream of coefficient whereby.
Alternatively, the designator in the block information table can be indicated from bitstream landmark (landmark) and be begun to move to the biased of this coefficient.Described graphical representation form can also comprise tabulation then, and this tabulation provides relevant each bitstream landmark to be arranged in the information of which image block.In this way, can move with the biased of position storage indicator of lesser number, this is because always move less than the biased of distance static position apart from biased the moving almost of bitstream landmark.This is hinting and can use less memory stores designator.As an alternative, need such tabulation, this tabulation provides each bitstream landmark to be arranged in the information of which image block.Therefore, when wanting specific image block of access, at first in this tabulation, check, to search the last bitstream landmark before this specific image block.Then, with relevant bitstream landmark and move relevant information apart from this boundary mark biased and be used for searching the desired images piece at this bit stream.
Designator can point to zeroth order or the coefficient of first order in the image block.When designator points to coefficient of zero order, the beginning of access data unit easily.Then, suffered even coefficient of zero order may be already be stored in block information table with the form of non-difference, the coefficient of zero order that also needs to decode is so that the access non-zeroth coefficients.When designator points to coefficient of first order, can the direct access non-zeroth coefficients.
In addition, the information of the bits number of indication between the coefficient of adjacent data cell, in the bit stream can be indicated the bits number between any combination of zeroth order and coefficient of first order.This information can be used for jumping to adjacent data cell at bit stream apace from a data unit.As mentioned above, by the access coefficient of zero order, perhaps can when learning, block information table pass through direct access coefficient of first order, the easily information of access data unit when coefficient of zero order.Therefore, the information of indicating bit number can be indicated to the bits number of coefficient of zero order or can be indicated to the bits number of coefficient of first order.Similarly, construct image presentation format by this way is so that it is suitable for indicating the bits number apart from coefficient of zero order or coefficient of first order.
In addition, designator can point to the coefficient in any data cell in the image block.Then, the information of the bits number between adjacent data cell can be used for any data cell of access image block.In addition, designator coefficient of zero order pointed, in the data cell is preferably and is stored in the block information table.Can use the information of the difference between the coefficient of zero order that is stored in the bit stream, calculate the coefficient of zero order of other data cell of this component.Zeroth order or coefficient of first order in first data cell of designator sensing image block are suitable.Therefore, by means of the directly beginning of access image block of designator.
Block information table can comprise and be used for being represented as and coefficient of zero order each coefficient of zero order of the difference of the coefficient of zero order of previous image block, that represent with non-differential form at bit stream.This is hinting each image block of access independently, and this provides all to be present in the bit stream, depended on the information of previous image block because of block information table.According to specific embodiment, block information table comprises the biased designator that move of indication from the bitstream landmark to the coefficient, and only with non-differential form, be used for being represented as coefficient of zero order with the coefficient of the difference of the coefficient of zero order of previous image block at bit stream.This embodiment provides has undersized block information table, and this is favourable under the memory capacity condition of limited.
Alternatively, block information table comprises the coefficient of zero order that each is represented with non-differential form.This is hinting does not need to be that data cell calculates coefficient of zero order, wherein the coefficient of zero order of these data cells in bit stream, be represented as with image block in the difference of coefficient of zero order of data cell.Therefore, the information in the access image block more quickly.Yet, in block information table, need more information.
In addition, bit stream can be represented digital picture with jpeg format.Therefore, the jpeg image of common JPEG compression or compression already can be associated with extra information, so that access, operation and/or analyze the specific part of this image apace.
According to the embodiment of the method for first aspect present invention, this method also comprises the image block that presents decoding to data processing unit or data representation unit, presents the part of image or image whereby with the ratio of dwindling.Data processing unit or data representation unit can or be used for the hardware cell of carries out image processing for for example screen, printer.A part that can be apace on screen for example, presents image or image, this be because this method with the relevant portion of mode decoding digital image very fast, so that on screen, present.This is hinting and can present image and not let user experiencing the tedious stand-by period to the user.
This method can also comprise carries out the calculating that is used for image processing to the image block of decoding.In this way, need less calculating, this is because this image is decoded as the simplification collection of huffman-coded coefficients.Therefore, carries out image processing relatively apace.
This method can also be included in to carry out and present performed result calculated when calculating on screen.This method provides the image that presents in real time with the compressed format storage when choosing image, and present operation in real time to image, this is because this method provides a kind of relevant portion of access graph picture and they are decoded as the method very fast of low volume data, so that can carry out and present operation apace.In addition, this method allows definition and carries out image operation on equipment such as mobile phone, that have little memory space.
According to the embodiment of the method for first aspect present invention, the number of coefficients by Hofmann decoding of data cell be used for approximate coefficient with greater number corresponding, through the image block of decoding.The coefficient that should be similar to the permission Hofmann decoding is represented more coefficient, can create whereby to have low-qualityer image, perhaps can create the less image that each image block has less pixel.Though this approximation loses the fine detail of image, the quality of image remains gratifying, this is because the coarse features of image is represented by first coefficient in each data cell.
By the predetermined number of the coefficient of Hofmann decoding can be for example four, nine, 13,18 or 24.When decoding four or 24 coefficients respectively, the coefficient of these decodings is represented the convergent-divergent of each image block from 8 * 8 pixels to 2 * 2 pixels and 4 * 4 pixels respectively.When Hofmann decoding nine, 13 or 18 coefficients, the coefficient of being decoded can be used for the expression of the image block of approximate 4 * 4 pixels.The number of these desorption coefficients is especially suitable, this be because adjacent the 9th, the 13 and the 18 coefficient after coefficient may not be used as the information of 4 * 4 pixel image block.
According to the embodiment of the graphical representation form of second aspect present invention, in block information table in a stream storage indicator, be used in reference to the information of the bits number between the coefficient that is shown in adjacent data cell and the coefficient of zero order of representing with non-differential form.In alternative embodiment, designator, coefficient value and image information are stored in the independent memory block.
According to the embodiment of the method for fourth aspect present invention, each checked bit sequence comprises sixteen bit.This form for bitstream entries is especially to be suitable for.The bitstream entries of coding non-zeroth coefficients comprises two parts.First is by huffman coding, and the zero run length of the current coefficient value of encoding and category.Second portion is the data of the current coefficient value of expression.Therefore the first of bitstream entries maintains how many positions how many coefficients are comprised by these clauses and subclauses codings (zero run length+1 coefficient) and this bitstream entries information.Because it is long that the first of bitstream entries is at most sixteen bit, so check that sixteen bit is suitable.Therefore, when once checking sixteen bit, each bit sequence will always comprise the first of a bitstream entries at least, and therefore this sixteen bit will keep the information of the length of relevant number of coefficients of being encoded by this bitstream entries and this bitstream entries.
Carry out the first eight bits that table lookup can comprise bit sequence and carry out first table lookup.If once carry out table lookup for sixteen bit, then need to have the form of 65,536 clauses and subclauses, this has expended very many memories.In addition, the first of most of common bitstream entries is eight or shorter.Therefore, carry out the information of in most of the cases needed, the relevant bitstream entries of table lookup of first eight bits with providing.
First table lookup can return the bit length of first bitstream entries and the information of the number of coefficients progressively passed through, perhaps turns back to the information of second table lookup.In this way, first table lookup will return relevant bitstream entries, needed information, perhaps return back eight information of analyzing the first of bitstream entries that are used to use this bit sequence.
Carry out table lookup and can also comprise second back eight table lookup that carries out bit sequence, with the bit length of determining first bitstream entries and the number of coefficients of progressively being passed through.When needs second table lookup, first table lookup can return sensing will be in second searches the pointer of form that use, that depend on first eight bits.This only needs several different forms to be used for second table lookup, and this is because have only the combination of the first eight bits of minority when being longer than eight in the first of bitstream entries.Therefore, when carrying out table lookup in two steps, do not need and the as many table entries of first bitstream entries that is used to analyze each bit sequence.
Embodiment according to the method for fifth aspect present invention uses from some image blocks in digital picture of information operating of other digital picture.This operation can comprise: if the content of image block is imaged in two digital pictures, then mix the influence from the image block of these two digital pictures.This is hinting this two digital pictures amalgamation each other more glibly.
According to further embodiment, at first handle the image block of the part of first digital picture, but temporarily store the image block of the remainder of this image with not compressed form.Then, the image block of this remainder can be used for calculating the operation to second digital picture before the image block of handling second digital picture.
In addition, the image block of the designator access amalgamation sequentially image by using the graphical representation form, and storing the huffman-coded coefficients stream that is used for the consecutive image piece with being represented as with the coefficient of zero order of the difference of corresponding previous coefficient of zero order, can be another image compression format such as jpeg format with the image transitions of institute's amalgamation.Therefore, can amalgamation and the operand word image need little memory span simultaneously, and when amalgamation is finished, this image is in the presentation format of being compressed better.
Description of drawings
Referring now to accompanying drawing the present invention is described in more detail by way of example, in the accompanying drawings:
Fig. 1 is the flow chart that is used for the method for compressed digital video according to embodiments of the invention.
Fig. 2 a-c is the schematic diagram according to the graphical representation form of the embodiment of the invention.
Fig. 3 is a flow chart of analyzing the method for the data cell in the bit stream.
Fig. 4 is the flow chart that is used to read with the method for the specific part of the digital picture of the image file format storage of compression.
Fig. 5 is according to embodiments of the invention, is used for decoded data units so that operate the flow chart of the method for the digital picture of storing with the image file format of compression.
Fig. 6 is according to embodiments of the invention, is used for the flow chart of two digital picture amalgamations for the method for the image file format of compression.
Fig. 7 is the schematic overview of two digital pictures of amalgamation.
Fig. 8-the 9th obtains digital picture and with the screenshot capture of equipment of digital picture amalgamation for the image file format of compression.
Embodiment
In the following description, though it is also conceivable that other compression of using other conversion, the compression of image will be described with reference to the JPEG compression.Should be noted that protection scope of the present invention is in no way limited to the JPEG compression.
Referring now to Fig. 1,, will a kind of method that is used for compressed digital video be described.At first, in step 10, represent digital picture with the YUV color model, wherein each pixel in the image has three components: brightness Y, and two chromatic component U and V.Two chromatic component U and V represent color characteristic, and their fine detail is difficult to be discovered by human eye.Therefore, these components can be represented with the resolution that is lower than luminance component.As 16 * 16 block of pixels of image of the image block example of definition herein, can by four each be that the V data cell of the U data cell of the Y data cell of 8 * 8 pixels, 8 * 8 pixels and one 8 * 8 pixel is represented.Represent to compare with the RGB of digital picture, this is corresponding to 50% compression of data set.Yet, can also use four U data cells and four V data cells, wherein represent to compare with RGB, do not obtain compression.
This image is treated to 16 * 16 discrete image blocks with any order.In step 12, use each data cell in each component of discrete cosine transform (DCT) conversion.Because each data cell of each component comprises 64 pixels, so DCT will generate 64 coefficients for the basic function of DCT.These coefficients comprise a coefficient of zero order (DC coefficient) and 63 coefficients of higher order (AC coefficient).
Next, in step 14, create the thresholding and the quantization approximation of the coefficient of each data cell.By each coefficient is come convergent-divergent and each coefficient of intercepting divided by a value according to normalization matrix, and realize thresholding and quantization approximation.This is hinting that the coefficient to the basic function that is confirmed as having low perceptual importance gives low weight, and to a large amount of coefficient values 0 (zero).
Then, use the zigzag order that the coefficient rearrangement is flowed as coefficient.According to the JPEG of standard compression, the DC coefficient is represented as the difference with the previous DC coefficient of the plurality of previous data units of same color component, and this difference is by huffman coding.The AC coefficient is carried out zero run length encode and further encode, and it directly is stored in after the DC of huffman coding coefficient with huffman coding.When only residue has the AC coefficient of null value in data cell, in this coefficient stream, insert end-of-block code.Therefore, obtained the bit stream of the huffman-coded coefficients of continuous data cell and image block.According to embodiments of the invention,, calculate coefficient through huffman coding in similar mode with regard to standard JPEG compression in step 16.This bit stream has constituted the compression expression of digital picture.
Yet, because used zero run length coding and huffman coding, so the length of each data cell is unknown.Therefore, before bit stream was decoded to the beginning of data cell from the beginning of this bit stream, the beginning of data cell was unknown.In addition, because previous DC coefficient is depended in the expression of DC coefficient, so if also do not determine previous DC coefficient by the huffman-coded stream of the previous coefficient of decoding, then the DC coefficient of data cell is unknown.
In order to allow specific in the rapidly retrieving images and to operate and/or analyze this specific thus, in step 18, create block information table, it covers the DC coefficient of each color component of the designator of each image block, the information of indicating the bits number in the bit stream that has between the adjacent coefficient of specifying order and each image block, and wherein the DC coefficient is represented with non-differential form.
According to other embodiments of the invention, the bit stream of the coefficient of huffman coding can have different contents or can differently be stored.For example, because depend on that the DC coefficient of previous image block is stored in block information table already and has suffered, so they do not need to be present in the bit stream.
In addition, all information of can be from this block information table known relevant image block are comprising the designator of DC coefficient that depends on other image block and the image block in the bit stream.Therefore, this bit stream does not need to be represented as the particular sequence of image block, does not perhaps even need to be stored in the stream.
Therefore, according to one embodiment of present invention, information specifies the mode of what part of its presentation video to be stored in the block information table with the position of a kind of wherein block message in form.Block message can also be stored in the mode that store with other block message its represented position wherein.Be used for image block coefficient huffman-coded stream in addition can combined block information store together.The block message of being stored can also have and is used for the mechanism of determining whether piece has been encoded, and this makes and the piece that is not encoded might be treated as for example black patch.
Therefore, the huffman-coded stream of coefficient and the encapsulation of block information table have constituted the digital image representation that needs little memory capacity, and it allows the specific part of analysis and application drawing picture and does not need the entire image of decoding simultaneously.This also allows with the nonlinear way compressed image, and wherein the compressed order of piece is not substantial.
Referring now to Fig. 2 a,, provided the structure of the graphical representation form of compression.Block information table can be stored among the RAM of equipment, and the arbitrary access that is used to read or write image block is that this equipment is needed.Shown in Fig. 2 b, block information table 30 and can be stored in the independent memory space through the bit stream 36 of the coefficient of huffman coding.Block information table 30 comprises designator 31, the information 32 of bits number of indication between the zeroth order or coefficient of first order of adjacent data cell, in the bit stream, and with the DC coefficient 33 of non-differential form storage.Therefore, the designator 31 in the block information table 30 covers the indication with first coefficient of specifying order of the image block in the bit stream 36.
As the substitute mode shown in Fig. 2 c, block information table 30 ' can be stored as bit stream 36 ' head or to bit stream 36 ' special sign, this bit stream 36 ' can constitute the jpeg format of standard.
In addition, should recognize that when in the head of information stores at bit stream, this head does not need to comprise needed full detail in the block information table.As an alternative, this head can comprise the information that is used to allow to create very apace block information table.Bit stream 36 ' head therefore can only keep the length information of each data cell.Then, when access graph as the time, can create block information table apace, and it is loaded among the RAM.Use the length of data cell, apace needed information in the access block information form.At the designator of the beginning that is created to each image block, decoding and storage DC coefficient and upgrade during the information that is used in reference to the bits number in the bit stream between the zeroth order that is shown in the adjacent data cell or the coefficient of first order, progressively pass through data cell.
This head even can be compressed is for example by huffman coding, so that save memory space.Length information has the value that only changes a little.Therefore, length is carried out huffman coding, incite somebody to action compressed information fully with the corresponding mode of encoding D C coefficient.The abundant compression of file is hinting that this document can allow to create apace block information table simultaneously with the effective form storage in space.Then when access graph as the time this head of decoding.
Yet, should be noted that the present invention is in no way limited to the mode of these presentation videos that provide.
Alternatively, designator 31 can be indicated biased the moving from the bitstream landmark to the coefficient.If designator 31 is indicated biased the moving that begins from the beginning of bit stream, then need be by the designator of 4 byte representations greater than the bit stream of 2 Mbytes sizes.Bitstream landmark can be inserted in the bit stream, so that make the designator 31 can be by 2 byte representations.When skew becomes greater than for example 65536 (by 2 maximum numbers that byte is represented), the registration bitstream landmark.Bitstream landmark is sequentially numbered.Create the tabulation of bitstream landmark, it provides relevant each bitstream landmark to be arranged in the information of which image block.In the time will using this designator access image block, access is carried out in two steps.At first, in bitstream landmark tabulation, compare, to search before will be, to be arranged in the numbering of the bitstream landmark of the approximated position of bit stream by the image block of access.Then, biased move of this image block apart from the bit stream beginning can be calculated as: biased the moving that (numbering of bitstream landmark) * 65536+ is provided by this designator.Certainly, between bitstream landmark, can use the position of arbitrary number.
The information of the bits number of indication between the adjacent coefficient of specifying order can be used for the particular-data unit of access image block apace.The position that this previous data cell that do not need to decode learns that data cell begins.This can be advantageously used in as described in greater detail below and quickly decode image blocks to the simplification collection of desorption coefficient.
Block information table can comprise and be used for DC coefficient each data cell, that represent with non-differential form.This shows does not need to calculate the DC coefficient for any data cell.Yet alternatively, block information table can only be included in those DC coefficients that are represented as in the bit stream with the difference of the DC coefficient of previous image block.When image block comprises several data cell for each component, only need to store the DC coefficient of first data cell of each component.This is hinting that block information table needs memory space still less.
Also can begin the structure of the graphical representation form that obtains to compress by jpeg file from compression already.Then, decode this jpeg file so that be determined to the designator of each image block, the information of the bits number of indication between the adjacent coefficient of specifying order and the DC coefficient that is used for each component of each image block with non-differential form.Therefore, in block information table, the huffman-coded stream of simultaneity factor is kept intact with index and DC coefficient storage.Alternatively, the AC coefficient of such jpeg image can be copied in the new memory space, full detail that whereby can the access jpeg image and not needing keeps the original jpeg image in the equipment of being stored in.
The graphical representation form of Chuan Jianing has been opened direct possibility to the jpeg file carries out image processing like this.
To describe now and a kind ofly be used to analyze jpeg image so that create the ad hoc approach of this graphical representation form.This method comprises sequentially progressively by bit stream so that collect needed information.In progressively by bit stream, will arrive in the designator, indication bit stream of each image block DC coefficient storage at least the first data cell of the information of specifying the bits number between the adjacent coefficient of order and each component in block information table.
Progressively by bit stream the time, the DC coefficient of decoding component, and temporarily store last coefficient so that make it possible to determine next DC coefficient.
Referring to Fig. 3, will the process of progressively passing through the AC coefficient of the data cell in the bit stream be described.The position that this analysis only wants definite next data cell to begin, and the bits number between the DC of an adjacent data cell coefficient or an AC coefficient.Therefore, do not need the AC coefficient of decoding.
The bitstream entries of AC coefficient comprises two parts.First is in the Huffman code that has how many zero coefficients and current coefficient value to encode in which category before the current coefficient.This category is determined the bits number of the second portion of bitstream entries, and its actual value to coefficient is encoded.Therefore, by analyzing the first of bitstream entries, can determine by the number of coefficients of this bitstream entries coding and the bits number of using by this bitstream entries.
In step 20,, comprise the bit sequence of checking the position of predetermined number in the bit stream progressively by the AC coefficient.Be preferably, bit sequence is that sixteen bit is long, and it is corresponding to the longest first of any bitstream entries.Therefore, by once checking sixteen bit, always can collect the information needed that is used at least one bitstream entries.Collect this information by in step 22, carrying out table lookup.Therefore this table lookup returns the bit length of at least the first bitstream entries that exists in this bit sequence, and by the number of coefficients of Huffman code coding.Then, in step 24, skip a plurality of positions accordingly with determined bit length.In step 26, the accumulation number of the position of being skipped is sued for peace, and in step 28, the accumulation number of the coefficient skipped is sued for peace.If also do not run into the block end symbol but also do not skip the coefficient of the maximum number of data cell, then this processing turns back to step 20, and checks new bit sequence.
Table lookup in step 22 can be carried out in two steps.At first, the first eight bits in this bit sequence is carried out table lookup.For the major part that is used for common jpeg image was searched, this searched the first that is enough to determine Huffman code.If first table lookup is enough to determine the first of first Huffman code, then no longer further analyze back eight this moment.If the first of Huffman code is longer than eight, then need further table lookup.First table lookup returns the pointer that points to new form then, will carry out back eight searching in this new form.This pointer depends on the first eight bits in the bit sequence.Second table lookup will be determined the bits number of Huffman code and the number of coefficients of skipping then.Second table lookup can carry out in different forms or in the different piece of identical form.The first eight bits of being longer than eight Huffman code only has the variant of minority.Therefore, the number of entries in second form is limited, and significantly less than 65536 (that is the sums of possible sixteen bit sequences).
Alternatively, the table lookup in the step 22 is carried out in single step.Then, be that 16 bit code executive table are searched.This table lookup sometimes may return the information of two or more Huffman codes.Therefore, when bit sequence comprises the first of several Huffman codes, can skip these Huffman codes simultaneously.When advancing row-octet and search, also might skip several Huffman codes simultaneously.Yet, have only the minority Huffman code short as to be enough to provide needed information with eight positions.
Referring now to Fig. 4,, will a kind of method that reads with the specific part in the digital picture of described graphical representation form storage be described.
At first, in step 40, position or the zone paid close attention in the definition image.Then, in step 42,, determine and the corresponding correct one or more image blocks in defined position or zone by simply displacement being associated with the image block sequence to the right and downwards respectively from the upper left corner.After this, in step 44,, determine the correct position of one or more image block in bit stream by in block information table, searching.In addition, in step 46, the DC coefficient of one or more image block that retrieval is correct from block information table.Next, in step 48, the described position in the access bit stream, and, use the DC coefficient decoding of being retrieved at this locational image block in step 50.
Referring now to Fig. 5,, will a kind of method of handling and operating with the digital picture of this graphical representation form storage be described.The processing of this digital picture and operation are that consequently it can be carried out on mobile phone and show in real time to the user in real time so fast.Therefore, the user can define the operation that will carry out, and sees that they were performed in several seconds.
At first, the user can define the image that will show or the part of image.Decoded picture only is the simplification collection of each data cell decoding through the coefficient of huffman coding so that it is presented on the screen in order to allow apace.The resolution approximate image that the simplification collection of this coefficient can be used for reducing is perhaps used less pixel display image.The number of decoded AC coefficient may suitably be zero, 8 * 8 pixel image block are reduced to 1 pixel, perhaps suitably is four, so that this image block is reduced to 2 * 2 pixels, perhaps suitably is 24, so that this image block is reduced to 4 * 4 pixels.When image block was reduced to 4 * 4 pixels, the fewer purpose AC coefficient of can decoding was so that approximate this 4 * 4 pixel image block.Therefore, for example nine, 13 or 18 coefficients of can decoding.Because the main information of each image block is present in first coefficient, so the information of losing in not decoded coefficient is not very important.
Because this graphical representation form, can retrieve very apace and decode through the simplification collection of the coefficient of huffman coding, so that show the part of this image or image.At first, in step 60, determine the image block that to decode.In step 62, by using each image block of designator access in the block information table.In step 64, also provide the DC coefficient of first data cell of each component by block information table.Then, in step 66, the AC coefficient of decoding desired number.After this,, skip the remainder of AC coefficient, come the next data cell of retrieving images piece apace by using for example information of the bits number between the AC coefficient of indication in adjacent data cell in step 68.Now, the next data cell of can decoding.In this way, decoded picture is so that present this image on screen very apace, and this has reduced the stand-by period that makes the user dislike.When use such as mobile phone, when having the unit of reduction process ability and little memory space, this is particularly useful.
The user can define the operation that will carry out the image that presents on the screen then.Operating needed calculating can carry out on the simplification collection of the coefficient of Hofmann decoding now.Therefore, can carry out calculating more quickly, and display result on display in real time, and do not let user experiencing the long stand-by period.
Referring now to Fig. 6-9,, will be described a kind of being used for the method for two digital picture amalgamations for the image file format of compression.The flow chart of the general introduction that provides this method has been shown among Fig. 6.The schematic overview of amalgamation image has been shown among Fig. 7, and Fig. 8-9 shows when this method different step in this method when equipment such as the mobile phone with embedded camera, that be used for obtaining image is realized.
At first, in step 100, obtain first digital picture 200, the content of this digital picture is displayed in the view finder of camera of Fig. 8.Then the image that obtains of conversion in case consider lens correction and with the image projection of being proofreaied and correct on cylinder.Therefore, the image that is obtained is suitable for being stitched on another image, so that create panoramic picture.Then as described in, in step 102, with the graphical representation form of the part B boil down to compression of this digital picture 200, wherein according to the current locus of image block in image, be distributed in location index in this image to these image blocks with reference to figure 1.In step 104, the other parts A of memory image is with the mix operation that is used for will carrying out after a while.If the user wants to create panoramic picture, then can define direction of panorama at the equipment that is used for obtaining image.Then, a part that approaches direction of panorama in first digital picture of being obtained 200 most is stored as the part of storing for mixed purpose.
In step 106, in the time will obtaining second image, this part A of first image can be present in the view finder of camera.As shown in Figure 9, this part of first digital picture can be present in the ground floor on the view finder, be exactly transparent wherein, watch and be presented on object space in the second layer of view finder in the back of the described part of this first digital picture so that can feel by camera every a pixel.If along right-hand lay structure panorama, then the right half of first digital picture will be present in the Far Left part of view finder.If camera itself is carried out lens correction and conic projection in real time as mentioned above, then the A of storage area of first image can be presented in the view finder simply.Yet, if carrying out in real time, do not proofread and correct by camera, part A is inversely transformed into the leftmost position that is positioned at view finder, so that cooperate better with the object space shown in the view finder.Therefore, in step 108, the guiding user obtains second image 202, and it locatees the corresponding object in first and second digital pictures in the overlapping region of view finder.The image that obtains of conversion then, so as to consider lens correction and with the image projection of being proofreaied and correct on cylinder.
Second digital picture comprises portion C now, and it is in fact corresponding to the part A in first digital picture.Therefore, in step 110, camera can be easily associated with each other with two digital pictures, so that make this two images amalgamation correctly each other.Certainly, can obtain association between these two digital pictures with any alternate manner.For example, computer unit can calculate and find the association between these two images, and perhaps for example when not having the overlapping region, this association can be defined by the user.Image displacement is relative to each other determined in this association.
Next, in step 112,, the part A that is stored for mixing is mixed mutually with second image according to this association.Then, in step 114, as described in reference to figure 1, utilize the correct index of image block, according to the graphical representation form of determined displacement with the compression of the part D boil down to of second image.In addition,, store another part E of second image, with the purpose of the mix operation that is used in the time further image will being captured panorama, will carrying out with not compressed form in step 116.As shown in Figure 7, because two image that is obtained displacements relative to each other, so part E may comprise the part that does not keep image information.Therefore, the memory device of image section E has and is used for determining the whether mechanism of presentation video data or unknown message of the pixel of being stored that causes owing to displacement.This information is used by mixed process after a while, and can use in operation associated.
Then, in step 118, if desired, repeating step 106-116 so that with further image be stitched to amalgamation image in.As described in reference to figure 1, the image of Tian Jiaing is described graphical representation form by reduced overall at last.
Next, define the maximum rectangular image that may from two or more digital pictures, form, and determine on the upper left corner, to be suitable for fully first image block of this rectangle.Then, in step 120, the huffman-coded stream of the coefficient by the input continuous blocks, from determined BOB(beginning of block) and from left to right and from top to bottom move through determined maximum rectangular image, with this graphical representation format conversion jpeg image file that is the amalgamation image.In such a way, will represent the image of a big amalgamation with the jpeg image file form.The digital picture of institute's amalgamation may be displayed on the screen of the mobile phone that comprises camera now.
Be described as a sequence though will be used for the method for amalgamation, wherein at first with the image file format of first digital image compression for compression, and after a while second digital picture is added in the graphical representation form of this compression, but should consider, can be by creating the graphical representation form of this compression in the graphical representation form that these two digital pictures is directly merged to compression, perhaps as selection, two digital pictures can each all be represented with the image file format of compression, then via with the compression the graphical representation form expression and be combined as a big image.
Now, opposite with the amalgamation image, will a kind of method that is used for the image of cutting JPEG baseline encoded be described.At first, analyze this jpeg image by decoding through the data of huffman coding.During this decode procedure, be stored in the block information table to the designator of each image block and the information of the bits number of indication between data cell.In addition, calculate the DC coefficient of first data cell of each color component that is used for each image block, and it is stored in the block information table.Now, by the zone of at first determining from this cutting process, to detain, finish the expression that jpeg image is cut to the new JPEG baseline encoded of a part of image.For each leftmost image block of going, need to use the new difference of the DC coefficient information calculating DC coefficient in the block information table in this zone.After this, determine the bit length that will duplicate by using index in the block information table, can be simply with the expression of remainder in this row from the data bitwise copy of the huffman coding of original jpeg image to new jpeg image.Therefore, can create the jpeg image of new cutting easily.
If will carry out the not compressed expression that jpeg image is cut to image on the contrary, then use index in the block information table to determine image block in this zone, and this image block is decoded to the tram in the not compressed expression of this image.Use the DC coefficient in the block information table to carry out decoding, the data of the whole huffman coding that in this example, do not need whereby to decode.
In addition, when image information is represented as the DCT coefficient, also can operate jpeg image by handling this image.By jpeg image being converted to the graphical representation form that provides herein, can be to these image block executable operations when image block is represented as the DCT coefficient.Therefore, use matrix operation, can for example rotate or zoomed image.
What should emphatically point out is that preferred embodiment described herein is restrictive anything but, and many alternative embodiments all are possible in the defined protection range by claims.
Claims (26)
1, the method for the position that begins of a kind of next data cell that is used for determining the bit stream of image, described method comprises:
Analyze the first of bitstream entries, described bitstream entries is included in the data cell;
According to described analysis, determine the bits number of using by described bitstream entries to the described first of described bitstream entries;
Skip and be confirmed as the bits number that to use by described bitstream entries; And
Determine whether to arrive the block end symbol,, then analyze the new bitstream entries that begins in last bitstream entries end if also do not arrive the block end symbol.
2, the method for claim 1, the described first of wherein said bitstream entries comprises and is illustrated in the Huffman code of value in which category that how many zero coefficients and current coefficient are arranged before the current coefficient that described category is determined the bits number of the second portion of described bitstream entries.
3, method as claimed in claim 2, the bits number in the wherein said bitstream entries is determined by table lookup.
4, the method for claim 1 also comprises:
To the number summation of the position of being skipped, till running into the block end symbol.
5, the method for claim 1 also comprises:
The information of the bits number of storage indication between the adjacent coefficient of specifying order.
6, the method for claim 1 also comprises:
Storage indication the biased designator that moves from the beginning of bit stream to the coefficient of specifying order.
7, the method for claim 1, wherein said image is encoded according to Joint Photographic Experts Group.
8, method as claimed in claim 2 also comprises:
Check the bit sequence of the position that has predetermined number in the described bit stream, and
Wherein said analytical procedure comprises that the bit sequence to being checked carries out table lookup.
9, method as claimed in claim 8, the step of wherein said table lookup comprises carries out first table lookup to the first eight bits in the described bit sequence, if and this first table lookup is not enough to determine the first of Huffman code, then carry out further table lookup.
10, method as claimed in claim 8, wherein said table lookup returns the information of at least two Huffman codes, and described skips steps comprises the bits number of skipping at least two bitstream entries.
11, a kind of method that is used for raw image data is encoded to the digital image representation of compression, described method comprises:
Read the image block of described raw image data middle finger sizing with any order, and for each image block:
This image block is transformed to one or more data cells of one or more component, and expression coefficient, each data cell as basic function is created in described conversion;
Calculate the quantization approximation of described coefficient;
At least some quantization parameters are expressed as the coefficient stream of consecutive image piece;
Described coefficient stream is carried out huffman coding, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component;
The described huffman-coded stream of packing coefficient in bit stream;
In block information table, store the zeroth order of each image block in the bit stream or the designator of coefficient of first order into; And
The coefficient of zero order of at least one data cell of each component of storage in block information table, described coefficient of zero order is represented with non-differential form.
12, method as claimed in claim 11, the described designator in the wherein said block information table are stored as the biased designator that moves from static position to this coefficient.
13, method as claimed in claim 11, described designator in the wherein said block information table is stored as the biased designator that moves from bitstream landmark to this coefficient, and the graphical representation form also comprises tabulation, and this tabulation provides relevant each bitstream landmark to be arranged in the information of which image block.
14, method as claimed in claim 11, wherein storage is used for being represented as and coefficient of zero order each coefficient of zero order of the difference of the coefficient of zero order of previous image block, that represent with non-differential form at bit stream.
15, as claim 11 or 12 described methods, wherein each coefficient of zero order is stored as with non-differential form and represents.
16, method as claimed in claim 11, wherein said bit stream is represented digital picture with jpeg format.
17, a kind of method of analyzing the digital picture of JPEG compression, the digital picture of this JPEG compression is represented as bit stream, wherein said bit stream is represented the consecutive image piece, each piece comprises one or more components, each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of the coefficient of basic function, and wherein coefficient of zero order is represented as the difference with the previous coefficient of zero order of respective component, and described method comprises:
Sequentially progressively pass through described bit stream, and progressively by described bit stream the time:
In block information table, store the zeroth order of each image block or the designator of coefficient of first order into; And
The decoding coefficient of zero order, and in block information table the coefficient of zero order of at least one data cell of each component of storage, described coefficient of zero order represents with non-differential form,
Wherein, the non-zeroth coefficients of the data cell in the described bit stream is represented that by the bitstream entries sequence progressively the step by these non-zeroth coefficients comprises:
Check the bit sequence of the subsequent bit of predetermined number in the bit stream;
Carry out table lookup, with the category of determining at least the first bitstream entries in this bit sequence and zero run length and the bit length of determining first bitstream entries;
Skip in the bit stream and the corresponding bits number of determined bit length;
The number of coefficients of progressively having passed through is sued for peace, up to the whole coefficients that progressively passed through described data cell or till running into the block end symbol; And
To the summation of the bits number skipped, so as to collect relevant in adjacent data cell zeroth order or coefficient of first order between, the information of the bits number in the bit stream.
18, method as claimed in claim 17, wherein each checked bit sequence comprises sixteen bit.
19, method as claimed in claim 17, the step of wherein carrying out table lookup comprise the first eight bits of bit sequence and carry out first table lookup.
20, method as claimed in claim 17, wherein said first table lookup are returned the bit length of relevant described first bitstream entries and the information of the number of coefficients progressively passed through, perhaps turn back to the information of second table lookup.
21, method as claimed in claim 17, wherein, the step of carrying out table lookup comprises that also back eight to described bit sequence carry out second table lookup, with the bit length of determining described first bitstream entries and the number of coefficients of progressively being passed through.
22, method as claimed in claim 17, the described designator in the wherein said block information table are stored as the biased designator that moves from static position to this coefficient.
23, method as claimed in claim 17, described designator in the wherein said block information table is stored as the biased designator that moves from bitstream landmark to this coefficient, and the graphical representation form also comprises tabulation, and this tabulation provides relevant each bitstream landmark to be arranged in the information of which image block.
24, method as claimed in claim 17, wherein storage is used for being represented as and coefficient of zero order each coefficient of zero order of the difference of the coefficient of zero order of previous image block, that represent with non-differential form at bit stream.
25, method as claimed in claim 17, wherein each coefficient of zero order is stored as with non-differential form and represents.
26, a kind of method that is used for two digital pictures of amalgamation, described method comprises:
Determine the spatial relationship between described two digital pictures;
According to the described spatial relationship between described two digital pictures, give the image block allocation index of the digital image information in described two digital pictures;
Form the bit stream of expression consecutive image piece according to the index that distributed, wherein each piece comprises one or more components, and each component comprises one or more data cells, and each data cell is represented as the huffman-coded stream of the coefficient of basic function,
The image block information stores that will be used for each image block according to the position of image block is in block information table, and described image block information comprises:
The zeroth order of each image block in the described bit stream or the designator of coefficient of first order; And
The coefficient of zero order of at least one data cell of each component, described coefficient of zero order is represented with non-differential form.
Applications Claiming Priority (6)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| SE0303085-5 | 2003-11-18 | ||
| SE0303085A SE0303085D0 (en) | 2003-11-18 | 2003-11-18 | Method for creating a compressed digital image representation and image representation format |
| SE03030855 | 2003-11-18 | ||
| SE0303204-2 | 2003-11-27 | ||
| SE03032042 | 2003-11-27 | ||
| SE0303204A SE0303204D0 (en) | 2003-11-26 | 2003-11-27 | S method and device for image processing |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200480034015 Division CN100535940C (en) | 2003-11-18 | 2004-11-16 | Method for processing a digital image and image representation format |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101635853A true CN101635853A (en) | 2010-01-27 |
| CN101635853B CN101635853B (en) | 2011-11-23 |
Family
ID=29729109
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200480034015 Expired - Fee Related CN100535940C (en) | 2003-11-18 | 2004-11-16 | Method for processing a digital image and image representation format |
| CN 200910160228 Expired - Fee Related CN101635853B (en) | 2003-11-18 | 2004-11-16 | Method for processing a digital image |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200480034015 Expired - Fee Related CN100535940C (en) | 2003-11-18 | 2004-11-16 | Method for processing a digital image and image representation format |
Country Status (2)
| Country | Link |
|---|---|
| CN (2) | CN100535940C (en) |
| SE (1) | SE0303085D0 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108028928A (en) * | 2015-09-18 | 2018-05-11 | 皇家飞利浦有限公司 | For compression of images quickly and efficiently and the method and apparatus of decompression |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9137443B2 (en) | 2011-04-26 | 2015-09-15 | Blackberry Limited | Fast estimation of binary data length using memory corruption |
| CN105912289A (en) * | 2016-04-22 | 2016-08-31 | 青岛海信移动通信技术股份有限公司 | Startup LOGO image conversion and storage method and apparatus, and startup LOGO image displaying method and apparatus |
| CN108419086B (en) * | 2018-05-08 | 2019-04-05 | 北京理工大学 | A kind of method for compressing image based on huffman coding |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5408328A (en) * | 1992-03-23 | 1995-04-18 | Ricoh Corporation, California Research Center | Compressed image virtual editing system |
| AU5983794A (en) * | 1993-11-30 | 1995-06-19 | General Electric Company | Data processor for assembling transport data packets |
| CN100579230C (en) * | 1997-04-01 | 2010-01-06 | 索尼公司 | Image encoder and method thereof, picture decoder and method thereof |
| US6353682B2 (en) * | 1998-09-17 | 2002-03-05 | Xerox Corporation | Rotated read-out of JPEG compressed images |
| US6381371B1 (en) * | 1999-03-17 | 2002-04-30 | Hewlett-Packard Company | Method and apparatus for processing image files |
-
2003
- 2003-11-18 SE SE0303085A patent/SE0303085D0/en unknown
-
2004
- 2004-11-16 CN CN 200480034015 patent/CN100535940C/en not_active Expired - Fee Related
- 2004-11-16 CN CN 200910160228 patent/CN101635853B/en not_active Expired - Fee Related
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108028928A (en) * | 2015-09-18 | 2018-05-11 | 皇家飞利浦有限公司 | For compression of images quickly and efficiently and the method and apparatus of decompression |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1882966A (en) | 2006-12-20 |
| CN100535940C (en) | 2009-09-02 |
| CN101635853B (en) | 2011-11-23 |
| SE0303085D0 (en) | 2003-11-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7480418B2 (en) | Method for processing a digital image and image representation format | |
| US8769395B2 (en) | Layout objects as image layers | |
| US8036475B2 (en) | Compression for segmented images and other types of sideband information | |
| US8902992B2 (en) | Decoder for selectively decoding predetermined data units from a coded bit stream | |
| US8625912B2 (en) | JPEG 2000-like access using the JPM compound document file format | |
| EP0831657A2 (en) | Vertex-based hierarchical shape representation and coding method and apparatus | |
| CN111510643B (en) | System and method for splicing panoramic image and close-up image | |
| CN101635853B (en) | Method for processing a digital image | |
| US20090317007A1 (en) | Method and apparatus for processing a digital image |
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: SWEDEN MOBILE IMAGE CORP. Free format text: FORMER OWNER: SCALADO AB Effective date: 20120816 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20120816 Address after: Longde, Sweden Patentee after: Mobile Imaging In Sweden AB Address before: Longde, Sweden Patentee before: Scalado AB |
|
| CP02 | Change in the address of a patent holder |
Address after: Stockholm Patentee after: Mobile Imaging In Sweden AB Address before: Longde, Sweden Patentee before: Mobile Imaging In Sweden AB |
|
| CP02 | Change in the address of a patent holder | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20181109 Address after: Espoo, Finland Patentee after: NOKIA TECHNOLOGIES OY Address before: Stockholm Patentee before: Mobile Imaging In Sweden AB |
|
| TR01 | Transfer of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20111123 Termination date: 20191116 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |