CN1378180A - Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation - Google Patents
Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation Download PDFInfo
- Publication number
- CN1378180A CN1378180A CN 01110259 CN01110259A CN1378180A CN 1378180 A CN1378180 A CN 1378180A CN 01110259 CN01110259 CN 01110259 CN 01110259 A CN01110259 A CN 01110259A CN 1378180 A CN1378180 A CN 1378180A
- Authority
- CN
- China
- Prior art keywords
- value
- quantization table
- dispersive
- length
- bit stream
- 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.)
- Pending
Links
Images
Abstract
The data compressing method using dynamic quantization coefficient as base for dispersive cosine transformation includes the following steps: cutting the color amount into several blocks; converting the amount of different blocks into dispersive cosine transformation coefficients to generate the corresponding transformation coefficient set; quantizing each transformation coefficient set by means of quantization table to produce corresponding quantization value, set; coding each quantization value set by using varying length coding algorithm to produce bit flow; and regulating the quantization table in case that the bit flow length is not equal the present value to produce one new quantization valve set and coding the new quantization valve set until the length of the bit flow approaches the preset value. The present invention makes it possible to save memory space and raise drawing rate.
Description
The data compression method of the relevant a kind of graphics process of the present invention, particularly a kind of data compression method with dynamic quantization coefficiency as dispersive cosine switching foundation, can to the visual pattern structure of color value will be converted, be compressed into the identical archives of a plurality of sizes, data volume and enhancement drawing efficiency when storing to reach the minimizing graphic structure.
In three-dimensional (3-D) drawing course, the figure mapping is unusual important techniques, can utilize this technology to drawing to go out various scene, it is earlier many figures basic or commonly used to be stored in the storage unit, the part scene of drawing again according to the institute desire, obtain suitable figure by the drawing chip to storage unit, be attached to appropriate location, simultaneously by multiple functions such as the modification of drawing chip or fine settings, it is required to make the figure of attaching meet scene, like this, repeat the figure that the scene of each part is required and be attached to appropriate location one by one, can finish the drafting of a scene; As shown in Figure 1, it is the schematic flow sheet of figure access in the past, show among the figure that all figures 11 are to be stored in the storage unit 12, and enter in the database of area stores body 13, in the drawing course, needed figure can enter in 14 yuan of the Graphics Processing Unit in the drawing chip, to be for further processing.
Yet, in 3-D draws, for the scene that makes drafting does not fall within stiffly, dull, make the color transformed nature that seems, scene is vivid, the variation of each fraction all must be presented, this just needs considerable graph data to be stored in the storage unit, relatively, and when carrying out the access of figure, certainly will use the very many storage spaces and the frequency range of bus, drawing efficiency is low.
The object of the present invention is to provide a kind of data compression method with dynamic quantization coefficiency as dispersive cosine switching foundation, will convert the visual pattern structure of color value to, be compressed into the identical archives of a plurality of sizes, the data volume when storing to reduce graphic structure is also promoted drawing efficiency.
For achieving the above object, the present invention takes following technical measures:
Among the present invention, as shown in Figure 2, after all compressed steps 3 of figures 2 elder generations, be stored in again in the storage unit 5, and enter in the database of regional memory cell 6, and in the process of drawing, the desired compression figure can enter in the drawing chip, after decompression 7, enter in the Graphics Processing Unit 8, to be for further processing; Because mode of the present invention is after earlier all figures 2 all being compressed 3, in the storage unit of restoring 5, compress figures to the chip of drawing by storage unit 5 transmission during drawing, can save the space of a large amount of storage unit 5 and the frequency range of bus.
The present invention mainly is that each the color value that will convert in the structure of visual pattern of three kinds of colors (Y.U.V) value all is compressed into the identical archives of a plurality of sizes, data volume in the time of can reducing the graphic structure storage, and make the storage location of all files be systematicness, in order to the archive file that obtains required figure rapidly, get final product requiredly graphic structure via reduction, do not influencing under the figure image quality, can effectively save the effect of storage area and enhancement drawing speed.
The present invention is with the data compression method of dynamic quantization coefficiency as dispersive cosine switching foundation, it is the structure that will convert the visual pattern of three kinds of color values to, utilize the following step, each color value all is compressed into the identical archives of a plurality of sizes, is in regular turn: the color value is divided into most blocks; And utilize dispersive cosine conversion (DCT), convert the value of each block to the dispersive cosine conversion coefficient, and produce corresponding coefficient sets; Quantize each coefficient sets with quantization table again, and then produce the corresponding quantitative value set; Use change length coding algorithm (VLD), each quantized value set is encoded, produce bit stream simultaneously; And be greater than or less than preset value when the length of bit stream, and then adjust quantization table, produces a new quantized value set, and new quantized value gathered change the length coding algorithm, be equal to or less than and near preset value until the length of bit stream; Like this, can make the data storing after this graphic structure forms the compression of color value, and can make the size of each packed data all identical, not influence under the figure image quality, effectively save the effect of storage area and enhancement drawing speed to reach.
The inventive method is taked following concrete steps:
Data compression method with dynamic quantization coefficiency as dispersive cosine switching foundation of the present invention is to convert the structure of the visual pattern of three primary colours Y.U.V color value to, to be compressed into the identical archives of a plurality of sizes, to comprise the steps:
A. Y color value is divided into most blocks;
B. utilize dispersive cosine conversion, convert the value of each block to the dispersive cosine conversion coefficient, and produce corresponding dispersive cosine conversion coefficient set;
C. quantize each dispersive cosine conversion coefficient set with quantization table, and produce the corresponding quantitative value set;
D. use change length coding algorithm, each quantized value set is encoded, and produce bit stream;
E. be greater than or less than preset value when the length of bit stream, then adjust quantization table, produces a new quantized value set, and new quantized value gathered change the length coding algorithm, be equal to or less than and near preset value until the length of bit stream.
Wherein, described U, V color value also compress according to said method.
Wherein, described block is the block of 2n*2n, and n is a positive integer, the color value that the human eye resolution is high more, and the block n value of being cut apart is more little.
Wherein, described change length coding algorithm is the Huffman coding.
Wherein, the step of described adjustment quantization table is in regular turn:
Quantization table is divided into 10 groups, except that the DC value;
When the length of bit stream surpasses preset value, amplify in regular turn by group to the group of lowest number of maximum numbering, until producing a suitable quantization table, the length of bit stream is equal to or less than, again near preset value;
When the length of bit stream during less than preset value, dwindle in regular turn by the group of lowest number to the group of maximum numbering, until producing a suitable quantization table, the length of bit stream is equal to or less than, again near preset value.
Wherein, the step of described quantification dispersive cosine conversion coefficient set is in regular turn:
With Joint Photographic Experts Group, produce a default quantization table;
With each factor of this default quantization table, be adjusted to 2 close power value simultaneously, form one and simplify quantization table;
Utilize this simplification quantization table with the set of dispersive cosine conversion coefficient set being quantized into quantized value.
Reaching embodiment in conjunction with the accompanying drawings is described in detail as follows architectural feature of the present invention:
Fig. 1 is the schematic flow sheet of figure access in the past.
Fig. 2 is the used figure access schematic flow sheet of the present invention.
Fig. 3 is the step synoptic diagram of first embodiment of the invention.
Fig. 4 is the schematic flow sheet that first embodiment of the invention quantizes.
Fig. 5 is the schematic flow sheet that first embodiment of the invention is adjusted quantization table.
Fig. 6 is the schematic flow sheet of first embodiment of the invention.
Fig. 7 is that second embodiment of the invention U value is cut apart synoptic diagram.
Table 1 is the Joint Photographic Experts Group quantization table that the present invention first implements.
Table 2 is simplification quantization tables of first embodiment of the invention.
Table 3 is group's tables of first embodiment of the invention.
Table 4 is the quantization tables after first embodiment of the invention increases the tenth group.
Table 5 is the quantization tables after first embodiment of the invention increases the 9th group again.
Table 6 is the quantization tables after first embodiment of the invention downgrades first group.
As shown in Figure 3, it is the synoptic diagram of the inventive method first embodiment, at first carry out segmentation procedure 31, be after Structure Conversion with visual pattern becomes color space, and (this is existing technology to produce three kinds of color values of Y.U.V, so no longer add narration at this), Y color value is divided into a plurality of 8*8 blocks; And in dispersive cosine switch process 32, utilize the dispersive cosine conversion, convert the value of each 8*8 block to the dispersive cosine conversion coefficient, so that the value of each 8*8 block just distributes according to image frequency, and the image frequency that makes these blocks is by left-to-right and increase progressively from top to bottom, to form corresponding 8*8 dispersive cosine conversion coefficient; Then, by quantization step 33, quantize each 8*8 dispersive cosine conversion coefficient with quantization table, and produce corresponding 8*8 quantized value set, the upper left corner is the DC value in this set; In change length coding calculation (VLD) step 34, use the change length coding algorithm of Huffman coding again, each 8*8 quantized value is encoded, and produce bit stream; At last, adjust bitstream length step 35; When the length of bit stream is greater than or less than preset value, then adjust quantization table, produce a new quantized value set, and new quantized value gathered move the length coding algorithm, be equal to or less than and near preset value until the length of bit stream; And the color value of these U, V also can compress according to said method, so can will convert the structure of the visual pattern of three kinds of color values of Y.U.V to, is compressed into the identical archives of a plurality of sizes.
Shown in Fig. 4, table 1 and table 2, it is respectively the schematic flow sheet that first embodiment of the invention quantizes, the Joint Photographic Experts Group quantization table of first embodiment of the invention and the simplification quantization table of first embodiment of the invention; Among Fig. 4, quantization step 33 is to produce a Joint Photographic Experts Group quantization table 331 (as shown in table 1, this quantization table is 8*8 (row * row)) in regular turn; With each factor of this Joint Photographic Experts Group quantization table, be adjusted to 2 close power value simultaneously, simplify quantization table and form one, step 332 (as shown in table 2, this simplifies quantization table also is 8*8 (row * row)); Utilize this simplification quantization table that each 8*8 dispersive cosine conversion coefficient is quantized 333 again, and produce corresponding 8*8 quantized value set, step 334.
Shown in Fig. 5 and table 3,4,5,6, they are respectively the schematic flow sheet that the present invention first implements to adjust quantization table, group's table of first embodiment of the invention, quantization table after quantization table after quantization table after first embodiment of the invention increases the tenth group, first embodiment of the invention increase the 9th group again and first embodiment of the invention downgrade first group; Cooperate shown in Fig. 3 and the table 2, when the bitstream length that produces in the change length coding algorithm steps 34 is greater than or less than preset value, then in the step 35 of adjusting bitstream length, adjust quantization table, step 351, and the flow process of adjustment quantization table as shown in Figure 5, at first, carry out step 352, quantization table is divided into 10 groups (except that the DC value in the upper left corner, as shown in table 3); Carry out step 352, when the length of bit stream surpasses preset value, just increase maximum value of numbering group, the value of group's (promptly numbering 10) that is about to the maximum numbering of this quantization table is all amplified 2 times of (tables of comparisons 2 simultaneously, table 3 is with shown in the table 4), and produce new quantization table, step 354, carry out step 355 again, the value of new quantization table is changed the length coding calculation, when if formed bit stream still surpasses preset value, increase the value of Next Serial Number group again, it is step 356, the value of next one numbering (promptly numbering 9) group that also will this new quantization table is all amplified 2 times of (tables of comparisons 3 simultaneously, table 4 is with shown in the table 5), step 355, equally another the new quantization table (seeing step 354) that is produced is changed the length coding calculation, by that analogy, when the length of bit stream during greater than preset value, being amplified in regular turn to the group of lowest number by the group of maximum numbering (is respectively the step 353 shown in Fig. 5,356), change the length coding calculation up to the new quantization table that is produced, the bitstream length that is produced is equal to or less than and near preset value, promptly carry out step 357, store bit stream.
In addition, when originally bit stream during less than preset value, just downgrade lowest number group value, be step 353 ', the value of group's (promptly numbering 1) that is about to the lowest number of this quantization table is all dwindled 2 times of (tables of comparisons 2 simultaneously, table 3 is with shown in the table 6), and produce new quantization table, be step 354 ', value that again will this new quantization table changes the length coding calculation, be step 355 ', if formed bit stream still less than but when keeping off preset value, downgrade the value of Next Serial Number group again, step 356 ', also soon the value of the group of the next one of this quantization table numbering (promptly numbering 2) is all dwindled 2 times, and it is same with another the new quantization table that is produced, step 354 ', change the length coding calculation, step 355 ', by that analogy, when the length of bit stream during less than preset value, by the group of lowest number to the group of maximum numbering dwindle in regular turn (be respectively the step 353 shown in Fig. 5 ', 356 '), change length coding calculation, step 355 up to the new quantization table that is produced (step 354 ') ', the bitstream length of generation is equal to or less than and near preset value, carry out step 357 again, promptly store bit stream.
As shown in Figure 6, it is the process flow diagram of first embodiment of the invention, and method of the present invention is carried out segmentation procedure 31 earlier, and Y color value is divided into a plurality of 8*8 blocks; And in dispersive cosine switch process 32, utilize the dispersive cosine conversion to convert the value of each 8*8 block to the dispersive cosine conversion coefficient; And in quantization step 33, earlier the Joint Photographic Experts Group quantization table is adjusted to the simplification quantization table, and step 332 utilizes this simplification quantization table that each 8*8 dispersive cosine conversion coefficient is quantized again, step 333, and then produce a plurality of corresponding 8*8 quantized values; Then, carry out step 34, change the length coding calculation, use the change length coding calculation of Huffman coding, each 8*8 quantized value is encoded, and produce bit stream; At last, carry out step 35: adjust bitstream length; If bitstream length surpasses default, then increase the value of the maximum numbering group of this quantization table, step 353; Produce new quantization table, step 354; Carry out step 355 again, the value of this new quantization table is changed the length coding calculation, if formed bit stream is equal to or less than and near preset value, then this bit stream is stored step 357; But, when this bitstream length still surpasses preset value, the value of new quantization table Next Serial Number group (being that last group number deducts 1 group) must be increased, be step 356, change length coding calculation (step 355) up to the new quantization table (step 354) that is produced, the bitstream length of generation is equal to or less than and near preset value, then this bit stream is stored, be step 357, otherwise the value that continues to carry out new quantization table Next Serial Number group increase action, step 356; So can with aforementioned all surpass bit streams of preset value length, all be transformed into and be equal to or less than and store (step 357) near the equal length bit stream of preset value; In addition, when bit stream during less than preset value, then downgrade the value of the lowest number group of this quantization table, step 353 ', and produce new quantization table, step 354 ', value that again will this new quantization table changes length coding calculation, step 355 ', if formed bit stream is equal to or less than and near preset value, then this bit stream is stored, i.e. step 357; But, this bitstream length still less than but when keeping off preset value, the value of new quantization table Next Serial Number group (being that last group number adds 1 group) must be downgraded, step 356 ', change the length coding calculation up to the new quantization table that is produced, be step 355 ', the bitstream length of generation is equal to or less than and near preset value, then this bit stream is stored, step 357, otherwise the value that continues to carry out new quantization table Next Serial Number group downgrades action, step 356 ', like this, can be with aforementioned all bit streams less than preset value length, all be transformed into and be equal to or less than and, and store step 357 near the equal length bit stream of preset value; And two kinds of color values of U and V also can compress according to above-mentioned flow process, and form the bit stream of a plurality of equal length; Like this, by above-mentioned method, visual pattern can be converted to the bit stream that equal length was all changed and be compressed into to three kinds of color values behind the color space, can store into the identical archives of a plurality of sizes.
In addition, when desire is used one of them figure, compression Y, U and the V value of this figure are read, should compress Y, U and V value again quantizes and counter-rotating dispersive cosine conversion (IDCT) through counter-rotating, the Y that can obtain reducing, U and V value, can form visual pattern via conversion again, more than be conventional art, and non-summary of the invention; Moreover, because Y, the U of each figure and V value all are compressed into the identical archives of size, and be stored, so that the storage location of all files is archives size at interval, be same intervals, according to the regularity of storage location, finding compression Y, U and the V value that needs fast, and then can promote drawing speed.
Because human eyes are to three kinds of color value Y, the resolution of U and V is not quite similar, resolution to the Y value is high than U and V value, so the Y value is compressed with the foregoing description method, and for human eye resolution lower U and V value, also can flow process as shown in Figure 6 compress, its process and a last embodiment are roughly the same, do not exist together is in segmentation procedure 31, as shown in Figure 7, shown in its synoptic diagram of cutting apart for second embodiment of the invention U value, be that U value 4 is divided into four big blocks 41, again each big block 41 is cut apart the sub-district that forms 16*16 and be divided into 32*32 block of cells 42 (among the figure only cut apart wherein a big block illustrate), again in each big block 41, to be one group of block 43 with 2*2 block of cells 42, can be divided into 8*8 group block 43 altogether, and calculate the average U value of each group in block 43, as each the block value in the 8*8 block, and make in each big block 41 is by 8*8 group block 43, be formed with identical 8*8 block with a last embodiment, and the value in this 8*8 block is the average U value in the corresponding group of block 43, like this, U value 4 is divided into four 8*8 blocks, then, also each block is carried out dispersive cosine switch process 32 in regular turn, quantization step 33, change length coding algorithm steps 34 and adjustment bitstream length step 35, these steps are all identical with a last embodiment, so no longer add narration at this; The V value also can be compressed according to above-mentioned flow process, can store into the identical archives of a plurality of sizes.
Therefore, the present invention can be according to the resolution of the mankind to the different color value, the color value that the human eye resolution is higher is divided into more a plurality of 8*8 blocks, in order to the process of changing, compressing, this color value can be done meticulousr processing, and the color value that resolution is lower in addition, be divided into less block, so that required storage area, compression back more dwindles, can under the condition that does not influence the figure image quality, effectively dwindle the storage area.
Compared with prior art, the present invention has following effect:
By the above embodiments as can be known, the present invention can be according to the resolution of human eye, most figures that will be used for the 3-D drawing are compressed into the identical archives of size, and the archival location that makes the area stores body that is stored in storage unit has regularity, so in drawing, one of them figure that the illustrator selects for use in database, can comply with the position of the archive file of this figure in this database simultaneously at this area stores body, and there is a regularity this position, the index that can make the area stores body is according to this regularity, make this index move to the tram rapidly, and fast the packed data of this figure is read, and through bus transfer to the chip of drawing, after decompressing, enter graphic element and be for further processing; So the present invention is not influencing under the figure quality, figure is compressed into the identical archives of size stores, can effectively save the storage area on the one hand and promote drawing speed, on the other hand, general drawing chip all can be supported the demoder of this method, can reduce the cost on the hardware.
Say that to sum up the present invention uses the compression method of dynamic quantization coefficient as dispersive cosine switching foundation, can effectively save the storage area and promote drawing speed.
Foregoing is to utilize embodiment that technical characterictic of the present invention is described, is not to be used to limit protection scope of the present invention, even there is the people to change slightly on the basis of the present invention's design, must belong in protection scope of the present invention.
Table 1:
| ????16 | ????11 | ????10 | ????16 | ????24 | ????40 | ????51 | ????61 |
| ????12 | ????12 | ????14 | ????19 | ????26 | ????58 | ????60 | ????55 |
| ????14 | ????13 | ????16 | ????24 | ????40 | ????57 | ????69 | ????56 |
| ????14 | ????17 | ????22 | ????29 | ????51 | ????87 | ????80 | ????62 |
| ????18 | ????22 | ????37 | ????56 | ????68 | ????109 | ????103 | ????77 |
| ????24 | ????35 | ????55 | ????64 | ????81 | ????104 | ????113 | ????92 |
| ????49 | ????64 | ????78 | ????87 | ????103 | ????121 | ????120 | ????101 |
| ????72 | ????92 | ????95 | ????98 | ????112 | ????100 | ????103 | ????99 |
Table 2:
| ??16 | ??8 | ??16 | ??16 | ??16 | ??32 | ??64 | ??64 |
| ??8 | ??8 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??32 | ??64 | ??64 | ??128 | ??128 | ??64 |
| ??32 | ??32 | ??64 | ??64 | ??64 | ??128 | ??128 | ??64 |
| ??64 | ??64 | ??64 | ??64 | ??128 | ??128 | ??128 | ??128 |
| ??64 | ??64 | ??64 | ??128 | ??128 | ??128 | ??128 | ??128 |
Table 3:
Table 4:
| ??16 | ??8 | ??16 | ??16 | ??16 | ??32 | ??64 | ??64 |
| ??8 | ??8 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??32 | ??64 | ??64 | ??128 | ??128 | ??128 |
| ??32 | ??32 | ??64 | ??64 | ??64 | ??128 | ??256 | ??128 |
| ??64 | ??64 | ??64 | ??64 | ??128 | ??256 | ??256 | ??256 |
| ??64 | ??64 | ??64 | ??128 | ??256 | ??256 | ??256 | ??256 |
Table 5:
| ??16 | ??8 | ??16 | ??16 | ??16 | ??32 | ??64 | ??64 |
| ??8 | ??8 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??32 | ??64 | ??64 | ??128 |
| ??16 | ??16 | ??16 | ??32 | ??64 | ??64 | ??128 | ??128 |
| ??16 | ??16 | ??32 | ??64 | ??64 | ??256 | ??256 | ??128 |
| ??32 | ??32 | ??64 | ??64 | ??128 | ??256 | ??256 | ??128 |
| ??64 | ??64 | ??64 | ??128 | ??256 | ??256 | ??256 | ??256 |
| ??64 | ??64 | ??128 | ??256 | ??256 | ??256 | ??256 | ??256 |
Table 6:
| ??16 | ??4 | ??16 | ??16 | ??16 | ??32 | ??64 | ??64 |
| ??4 | ??8 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??32 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??16 | ??32 | ??64 | ??64 | ??64 | ??64 |
| ??16 | ??16 | ??32 | ??64 | ??64 | ??128 | ??128 | ??64 |
| ??32 | ??32 | ??64 | ??64 | ??64 | ??128 | ??128 | ??64 |
| ??64 | ??64 | ??64 | ??64 | ??128 | ??128 | ??128 | ??128 |
| ??64 | ??64 | ??64 | ??128 | ??128 | ??128 | ??128 | ??128 |
Claims (6)
1. the data compression method with dynamic quantization coefficiency as dispersive cosine switching foundation is to convert the structure of the visual pattern of three primary colours Y.U.V color value to, to be compressed into the identical archives of a plurality of sizes, to comprise the steps:
A. Y color value is divided into most blocks;
B. utilize dispersive cosine conversion, convert the value of each block to the dispersive cosine conversion coefficient, and produce corresponding dispersive cosine conversion coefficient set;
C. quantize each dispersive cosine conversion coefficient set with quantization table, and produce the corresponding quantitative value set;
D. use change length coding algorithm, each quantized value set is encoded, and produce bit stream;
E. be greater than or less than preset value when the length of bit stream, then adjust quantization table, produces a new quantized value set, and new quantized value gathered change the length coding algorithm, be equal to or less than and near preset value until the length of bit stream.
2. data compression method as claimed in claim 1 is characterized in that, described U, V color value also compress according to said method.
3. data compression method as claimed in claim 1 or 2 is characterized in that, described block is a 2n*2n block, and n is a positive integer, the color value that the human eye resolution is high more, and the block n value of being cut apart is more little.
4. data compression method as claimed in claim 1 or 2 is characterized in that, described change length coding algorithm is the Huffman coding.
5. data compression method as claimed in claim 1 or 2 is characterized in that, the step of described adjustment quantization table is in regular turn:
Quantization table is divided into 10 groups, and the upper left corner is the DC value in set;
When the length of bit stream surpasses preset value, amplify in regular turn by group to the group of lowest number of maximum numbering, until producing a suitable quantization table, the length of bit stream is equal to or less than, again near preset value;
When the length of bit stream during less than preset value, dwindle in regular turn by the group of lowest number to the group of maximum numbering, until producing a suitable quantization table, the length of bit stream is equal to or less than, again near preset value.
6. data compression method as claimed in claim 1 or 2 is characterized in that, the step of described quantification dispersive cosine conversion coefficient set is in regular turn:
With Joint Photographic Experts Group, produce a default quantization table;
With each factor of this default quantization table, be adjusted to 2 close power value simultaneously, form one and simplify quantization table;
Utilize this simplification quantization table with the set of dispersive cosine conversion coefficient set being quantized into quantized value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 01110259 CN1378180A (en) | 2001-04-03 | 2001-04-03 | Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 01110259 CN1378180A (en) | 2001-04-03 | 2001-04-03 | Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1378180A true CN1378180A (en) | 2002-11-06 |
Family
ID=4658458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 01110259 Pending CN1378180A (en) | 2001-04-03 | 2001-04-03 | Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1378180A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1504877B (en) * | 2002-12-05 | 2010-04-28 | 佳能株式会社 | Incremental color transform creation |
-
2001
- 2001-04-03 CN CN 01110259 patent/CN1378180A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1504877B (en) * | 2002-12-05 | 2010-04-28 | 佳能株式会社 | Incremental color transform creation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1303820C (en) | Quality based image compression | |
| CN1186942C (en) | Variance based adaptive block size DCT image compression | |
| US5903669A (en) | Image processing apparatus and method | |
| CN1145363C (en) | Static picture and cartoon cooding quantization matrix | |
| US6882753B2 (en) | Adaptive quantization using code length in image compression | |
| CN1893659A (en) | DCT compression using Golomb-Rice coding | |
| CN1592412A (en) | Adaptive variable-length coding methods for image data | |
| CN1675932A (en) | System and method for rate-distortion optimized data partitioning for video coding using backward adapatation | |
| CN103581691B (en) | A kind of towards sparse coefficient efficiently can parallel image coding method | |
| CN1547708A (en) | A system and method for decoding digital image and audio data in a lossless manner | |
| CN101068352A (en) | Network image compressing method and system | |
| CN114143557B (en) | Low-complexity coding method for wavelet transformation high-frequency coefficient of video image | |
| CN1254112C (en) | Diveided image coding and decoding method with arbitrary shape region segmentation | |
| CN1378180A (en) | Data compressing method using dynamic quantization coefficiency as dispersive cosine switching foundation | |
| CN106331719A (en) | K-L transformation error space dividing based image data compression method | |
| CN200962655Y (en) | A motive image compression system for mobile device | |
| CN1825894A (en) | All phase cosine double orthogonal transformation and JPEG improving method | |
| JP3835273B2 (en) | Image encoding / decoding device | |
| CN1822050A (en) | Full phase anticosine double orthogonal transformation and its improving method for JPEG | |
| CN100379251C (en) | Encoding method for static image compression | |
| CN1191717C (en) | Method and device for reducing blocking artifact in the DCT domain | |
| TW483267B (en) | Compression method using dynamic quantitative coefficients as the basis of discrete cosine transformation | |
| CN102307303A (en) | Ternary-representation-based image predictive coding method | |
| CN1545320A (en) | Code rate distribution method for fine grain expansible video coding | |
| Bethel et al. | Optimal quantisation of the discrete cosine transform for image compression |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |