CN102893592B - Image downscaling method and device - Google Patents
Image downscaling method and device Download PDFInfo
- Publication number
- CN102893592B CN102893592B CN201080061138.2A CN201080061138A CN102893592B CN 102893592 B CN102893592 B CN 102893592B CN 201080061138 A CN201080061138 A CN 201080061138A CN 102893592 B CN102893592 B CN 102893592B
- Authority
- CN
- China
- Prior art keywords
- dot matrix
- pixel
- processing device
- medium processing
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 239000011159 matrix material Substances 0.000 claims abstract description 480
- 239000000203 mixture Substances 0.000 claims description 27
- 238000010586 diagram Methods 0.000 description 13
- 230000006870 function Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000000205 computational method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/387—Composing, repositioning or otherwise geometrically modifying originals
- H04N1/393—Enlarging or reducing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4023—Scaling of whole images or parts thereof, e.g. expanding or contracting based on decimating pixels or lines of pixels; based on inserting pixels or lines of pixels
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of image downscaling method, described method comprises: it is the picture signal of unit that medium processing device obtains with 8 × 8 dot matrix; Medium processing device is selected pixels point from current 8 × 8 dot matrix obtained, and the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in, wherein, N is positive integer; Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix; Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.The present invention also provides a kind of medium processing device and image processing system.
Description
Technical field
The present invention relates to image processing techniques, particularly relate to image downscaling method and device.
Background technology
Image scaling techniques is widely used in various medium processing device.After picture signal enters medium processing device, picture signal will be supplied to display device by medium processing device.Because the resolution of display device is often different from the resolution of signal source, so, after medium processing device often needs to use image scaling techniques to carry out convergent-divergent process to picture signal, again the picture signal after convergent-divergent is supplied to display device, wherein, image of the present invention comprises video image and picture.
Such as, more common high-definition image resolution is 1440 × 1080 now, if need by resolution be 1440 × 1080 high-definition image be presented at the display device that resolution is 640 × 480, then need by resolution be 1440 × 1080 image down 2.25 (horizontal direction) × 2.25 (vertical direction) doubly after, could be show in the display device of 640 × 480 in resolution.
At present, owing to using the reasons such as the demand of the user of display device, image needs with preview mode display in the display device sometimes.After preview mode also requires that the picture signal of acquisition is narrowed down to certain proportion by medium processing device, then the picture signal after reducing is supplied to display device.In actual applications, decoder carries out discrete cosine transform (DCT by by the picture signal compressed, Discrete Cosine Transform) decoding, be such as Motion Picture Experts Group (MPEG by the picture signal compressed, Moving Picture ExpertsGroup) picture signal of form, JPEG (joint photographic experts group) (JPEG, Joint Picture Expert Group) picture signal etc. of form, decoder will input to medium processing device through the decoded picture signal of DCT decoding technique, medium processing device processes through the decoded picture signal of DCT decoding technique, and the picture signal after process is inputed to display device display.
Decoder generally will by through the decoded picture signal of DCT decoding technique with 8 × 8 dot matrix be unit export, after medium processing device obtains and is the picture signal of unit with 8 × 8 dot matrix, generally want multiple 8 × 8 dot matrix of buffer memory, afterwards, then carry out reducing process with behavior unit.
But, in the above prior art, after medium processing device obtains and is the picture signal of unit with 8 × 8 dot matrix, before process is reduced to picture signal, need multiple 8 × 8 dot matrix of buffer memory, like this, medium processing device needs cache resources to carry out multiple 8 × 8 dot matrix of buffer memory, and such cache resources is no small resource overhead for medium processing device.In addition, reduce process for picture signal, traditional method is mostly adopt leggy interpolation, and concrete mode is the output multiply-add operation of multiple sampling point being obtained to a pixel, but also need to add anti-aliasing filter, it is such that to reduce processing mode comparatively complicated.
Summary of the invention
The invention provides image downscaling method and device, in order to when reducing process to the picture signal taking 8 × 8 dot matrix as unit, saving the cache resources of medium processing device.
The invention provides a kind of image downscaling method, be applicable to image preview mode, be applied to the image processing system comprising decoder and medium processing device, what medium processing device exported decoder is unit with 8 × 8 dot matrix picture signal reduces process, the described picture signal being unit with 8 × 8 dot matrix is the picture signal that decoder obtains after DCT decoding, and described method comprises: it is the picture signal of unit that medium processing device obtains with 8 × 8 dot matrix; Medium processing device is selected pixels point from current 8 × 8 dot matrix obtained, and the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in, wherein, N is positive integer; Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix; Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.
The present invention also provides a kind of medium processing device, be applicable to image preview mode, be applied to the image processing system comprising decoder and medium processing device, what medium processing device exported decoder is unit with 8 × 8 dot matrix picture signal reduces process, the described picture signal being unit with 8 × 8 dot matrix is the picture signal that decoder obtains after DCT decoding, described medium processing device comprises: obtain unit, is the picture signal of unit for obtaining with 8 × 8 dot matrix; Choose unit, for selected pixels point from current 8 × 8 dot matrix obtained, the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in by the pixel chosen, and wherein, N is positive integer; Arithmetic element, for carrying out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix; Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.
In the present invention, after medium processing device obtains and is the picture signal of unit with 8 × 8 dot matrix, the pixel for participating in the picture signal taking 8 × 8 dot matrix as unit to reduce N process is doubly chosen from current 8 × 8 dot matrix obtained, computing is carried out to the pixel chosen, export pixel or the intermediate object program of corresponding N × N dot matrix, after multiple 8 × 8 dot matrix of buffer memory, then need not carry out reducing process, save the cache resources needed for multiple 8 × 8 dot matrix of buffer memory.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is a kind of logical construction schematic diagram of the image processing system that the present invention applies;
Fig. 2 is the schematic diagram of 8 × 8 dot matrix;
Fig. 3 is the flow chart of a kind of image downscaling method of the present invention;
8 × 8 dot matrix that Fig. 4 A is the minification in one embodiment of the invention when being 2 are divided the schematic diagram of situation;
The basic templates schematic diagram that Fig. 4 B is the minification in one embodiment of the invention when being 2;
8 × 8 dot matrix that Fig. 5 A is the minification in one embodiment of the invention when being 4 are divided the schematic diagram of situation;
The basic templates schematic diagram that Fig. 5 B is the minification in one embodiment of the invention when being 4;
3 × 38 × 8 dot matrix that Fig. 6 A is the minification in one embodiment of the invention when being 3 are divided the schematic diagram of situation;
The basic templates schematic diagram that Fig. 6 B and Fig. 6 C is the minification of one embodiment of the invention when being 3;
The composition schematic diagram of 16 × 16 dot matrix that Fig. 7 is the minification in one embodiment of the invention when being 16;
The schematic diagram of multiple 12 × 12 dot matrix of multiple 8 × 8 dot matrix composition that Fig. 8 is the minification in one embodiment of the invention when being 12;
Fig. 9 is a kind of realization flow figure of image downscaling method of the present invention;
Figure 10 is the logical construction schematic diagram of a kind of medium processing device of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
First the image processing system of the present invention's application is introduced.
As shown in Figure 1, image processing system comprises decoder, medium processing device and display device.It is the picture signal of unit that decoder exports with 8 × 8 dot matrix, and medium processing device reduces process to the picture signal taking 8 × 8 dot matrix as unit, and output pixel point is to display device.
Introduce some professional knowledges that the present invention relates to below.
As shown in Figure 2,8 × 8 dot matrix comprise 64 pixels, and in 8 row, 8 column distributions, often row comprises 8 pixels, often row comprise 8 pixels, and the position of any one pixel in 8 × 8 dot matrix is decided by the row and column at this pixel place in 8 × 8 dot matrix.Suppose row in 8 × 8 dot matrix by i (i be not more than 8 positive integer) represent, arrange by j (j be not more than 8 positive integer) represent, any one pixel so in 8 × 8 dot matrix can represent with A (i, j).
In the present invention, medium processing device is when reception is the picture signal of unit with 8 × 8 dot matrix, be that unit receives with 8 × 8 dot matrix, namely, after receiving the 1st 8 × 8 dot matrix, receive the 2nd 8 × 8 dot matrix again, until after receiving all pixels of first 8 row, receive the 1st 8 × 8 dot matrix of second 8 row, the 2nd 8 × 8 dot matrix again, until after having accepted all pixels of second 8 row, receive the 1st 8 × 8 dot matrix of the 3rd 8 row, the 2nd 8 × 8 dot matrix again, by that analogy, until receive 8 × 8 all dot matrix.
In the present invention, reduce N process doubly to the picture signal taking 8 × 8 dot matrix as unit to refer to and reduce N doubly to picture signal respectively with vertical direction in the horizontal direction, namely, reduce N (horizontal direction) × N (vertical direction) doubly to the picture signal taking 8 × 8 dot matrix as unit, reducing N to the picture signal taking 8 × 8 dot matrix as unit is doubly a kind of mode being convenient to describe.Reducing N to picture signal doubly can think to image down N doubly, also can think to reduce N doubly to each 8 × 8 dot matrix of composition diagram picture, and wherein, N is positive integer.In the present invention, a pixel is exported for each N × N dot matrix correspondence.
The basic templates mentioned in the following examples of the present invention is the template of medium processing device for selected pixels point, specifies the pixel chosen from dot matrix in template.The design of template is not unique, and in other words, from dot matrix, choose which pixel is not unique, specifically can be designed according to actual needs by those skilled in the art.
In addition, for coloured image, the information of a pixel can such as, by multiple representation in components, R, G, B or Y, Cb, Cr etc.When carrying out computing to the pixel chosen, computing should be carried out with regard to each component of this pixel respectively.Suppose corresponding 3 components of a pixel, be respectively component 1, component 2 and component 3, when carrying out computing to the pixel chosen, component 1 computing tackling this pixel obtains the component 1 of output pixel point, in like manner, component 2 computing of this pixel is obtained to the component 2 of the pixel exported, the component 3 computing of this pixel is obtained to the component 3 of the pixel exported, so just obtain 3 components of the pixel of output, also just obtain the pixel of output.
The all embodiment of the present invention is all applicable to image preview mode.The all embodiment of the present invention can be applied to the image processing system comprising decoder and medium processing device, what medium processing device exported decoder is unit with 8 × 8 dot matrix picture signal reduces process, and the picture signal being unit with 8 × 8 dot matrix is the picture signal that decoder obtains after DCT decoding.
Introduce a kind of image downscaling method of the present invention below, this method can describe from the angle of medium processing device.As shown in Figure 3, this method comprises:
S301: it is the picture signal of unit that medium processing device obtains with 8 × 8 dot matrix.
S302: medium processing device is selected pixels point from current 8 × 8 dot matrix obtained, the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in, wherein, N is positive integer.
S303: medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix; Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.
When medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained, mean operation can be carried out to the pixel chosen.
If medium processing device obtains the intermediate object program of corresponding N × N dot matrix, so medium processing device can store described intermediate object program.Medium processing device, after all intermediate object programs obtaining corresponding N × N dot matrix, carries out computing to all intermediate object program, can obtain the pixel of corresponding N × N dot matrix.
When medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained, summation operation can be carried out to the pixel chosen from current 8 × 8 dot matrix obtained.In this case, the intermediate object program of described corresponding N × N dot matrix is medium processing device carries out summation operation result to the pixel chosen from current 8 × 8 dot matrix obtained.When medium processing device carries out computing to all intermediate object program, mean operation can be carried out to the result of all summation operation, thus obtain the pixel of corresponding N × N dot matrix.Concrete, if N is the approximate number of 8, so 8 × 8 dot matrix can be divided into (8/N) (horizontal direction) × individual N × N dot matrix of (8/N) (vertical direction).
In this case, medium processing device can selected pixels point from each N × N dot matrix current 8 × 8 dot matrix obtained respectively, and respectively computing is carried out to the pixel chosen from each N × N dot matrix in current 8 × 8 dot matrix obtained, obtain the pixel of each N × N dot matrix in 8 × 8 dot matrix of corresponding current acquisition.
For N=2, medium processing device needs the picture signal received to reduce 2 times.As shown in Figure 4 A, 8 × 8 dot matrix can be divided into 4 × 42 × 2 dot matrix, and each 2 × 2 dot matrix export a pixel, and the pixel of output is designated as O (i, j).In this embodiment, basic templates as shown in Figure 4 B, from Fig. 4 B, all pixels in each 2 × 2 dot matrix participate in the computing of output pixel point, compute mode is the mean value of the value of getting pixels all in basic templates, from 2 × 2 dot matrix, namely chooses all pixels do mean operation.4 × 4 pixels of 8 × 8 dot matrix outputs can be obtained thus successively, as follows respectively:
O(1，1)＝[A(1，1)+A(1，2)+A(2，1)+A(2，2)]/4
O(1，2)＝[A(1，3)+A(1，4)+A(2，3)+A(2，4)]/4
………·
O(4，3)＝[A(7，5)+A(7，6)+A(8，5)+A(8，6)]/4
O(4，4)＝[A(7，7)+A(7，8)+A(8，7)+A(8，8)]/4
Again for N=4, medium processing device needs the picture signal received to reduce 4 times.As shown in Figure 5A, 8 × 8 dot matrix can be divided into 2 × 24 × 4 dot matrix, and each 4 × 4 dot matrix export a pixel, and the pixel of output is designated as O (i, j).In this embodiment, as shown in Figure 5 B, the direct-shadow image vegetarian refreshments in each 4 × 4 dot matrix participates in the computing of output pixel point to basic templates, and compute mode is the mean value of the value of getting direct-shadow image vegetarian refreshments in basic templates.2 × 2 pixels of 8 × 8 dot matrix outputs can be obtained thus successively, as follows respectively:
O(1，1)＝[A(1，1)+A(1，3)+A(3，1)+A(3，3)]/4
O(1，2)＝[A(1，5)+A(1，7)+A(3，5)+A(3，7)]/4
O(2，1)＝[A(5，1)+A(5，3)+A(7，1)+A(7，3)]/4
O(2，2)＝[A(5，5)+A(5，7)+A(7，5)+A(7，7)]/4
If N be less than 8 and be not 8 approximate number, so 8 × 8 dot matrix can be divided at least one N × N dot matrix and multiple boundary pixel point, boundary pixel point in 8 × 8 dot matrix can form N × N dot matrix with the boundary pixel point in 8 × 8 adjacent dot matrix, wherein, here boundary pixel point refers to the pixel except the pixel that N × N dot matrix comprises in 8 × 8 dot matrix, in other words, together participate in the pixel in adjacent 8 × 8 dot matrix the pixel forming N × N dot matrix in 8 × 8 dot matrix.
In this case, medium processing device can selected pixels point from each N × N dot matrix current 8 × 8 dot matrix obtained respectively, and respectively computing is carried out to the pixel chosen from each N × N dot matrix in current 8 × 8 dot matrix obtained, obtain the pixel of each N × N dot matrix in 8 × 8 dot matrix of corresponding current acquisition respectively.If many for the quantity for the boundary pixel point forming described N × N dot matrix in adjacent 8 × 8 dot matrix of the number ratio of the boundary pixel point forming N × N dot matrix in 8 × 8 dot matrix of current acquisition, so medium processing device can from current 8 × 8 dot matrix obtained for form described N × N dot matrix boundary pixel point in selected pixels point, and carry out computing to from current 8 × 8 dot matrix obtained for forming the pixel chosen in the boundary pixel point of described N × N dot matrix, obtain the pixel of corresponding described N × N dot matrix.
For N=3, medium processing device needs the picture signal received to reduce 3 times.In this case, 8 × 8 dot matrix need output 2.67 × 2.67 pixels, but in actual applications, do not have the pixel of non-integer number.To this, the least common multiple getting 3 and 8 is 24, like this, and 3 × 38 × 8 dot matrix outputs, 8 × 8 pixels.As shown in Figure 6A, 3 × 38 × 8 dot matrix can be divided into 8 × 83 × 3 dot matrix, and each 3 × 3 dot matrix export a pixel, and the pixel of output is designated as O (i, j).
Here it should be noted that, pixel in pixel in 1st 8 × 8 dot matrix, the 2nd 8 × 8 dot matrix and the pixel in the 3rd 8 × 8 dot matrix belong to the pixel of same 8 row, in like manner, pixel in pixel in 4th 8 × 8 dot matrix, the 5th 8 × 8 dot matrix and the pixel in the 6th 8 × 8 dot matrix belong to the pixel of same 8 row, and the pixel in the pixel in the 7th 8 × 8 dot matrix, the 8th 8 × 8 dot matrix and the pixel in the 9th 8 × 8 dot matrix belong to the pixel of same 8 row.Adjacent 8 × 8 dot matrix of the 1st 8 × 8 dot matrix are the 2nd 8 × 8 dot matrix and the 4th 8 × 8 dot matrix respectively, adjacent 8 × 8 dot matrix of the 5th 8 × 8 dot matrix are the 1st 8 × 8 dot matrix, the 2nd 8 × 8 dot matrix, the 3rd 8 × 8 dot matrix, the 4th 8 × 8 dot matrix, the 6th 8 × 8 dot matrix, the 7th 8 × 8 dot matrix, the 8th 8 × 8 dot matrix and the 9th 8 × 8 dot matrix respectively, adjacent 8 × 8 dot matrix of other 8 × 8 dot matrix can be determined according to this, repeat no more here.
Also it should be noted that, if the pixel of the image that medium processing device receives is more than 24 row, so medium processing device is accepting the 1st 8 × 8 dot matrix respectively, after 2nd 8 × 8 dot matrix and the 3rd 8 × 8 dot matrix, also need other 8 × 8 dot matrix receiving same a line successively, afterwards, just can receive the 4th 8 × 8 dot matrix successively, 5th 8 × 8 dot matrix and the 6th 8 × 8 dot matrix, in like manner, accepting the 4th 8 × 8 dot matrix respectively, after 5th 8 × 8 dot matrix and the 6th 8 × 8 dot matrix, also need to receive successively and these 38 × 8 dot matrix other 8 × 8 dot matrix with a line, afterwards, just can receive the 7th 8 × 8 dot matrix successively, 8th 8 × 8 dot matrix and the 9th 8 × 8 dot matrix.
In this embodiment, the basic templates of the 1st 8 × 8 dot matrix respectively as shown in Fig. 6 B and Fig. 6 C, the 1st 8 × 8 dot matrix output 3 × 3 pixels, 3 × 3 pixels are followed successively by:
O(1，1)＝[A(2，1)+A(2，3)]/2
O(1，2)＝[A(2，4)+A(2，6)]/2
O(1，3)＝[A(2，7)+A(2，8)]/2
O(2，1)＝[A(5，1)+A(5，3)]/2
O(2，2)＝[A(5，4)+A(5，6)]/2
O(2，3)＝[A(5，7)+A(5，8)]/2
O(3，1)＝[A(8，1)+A(8，3)]/2
O(3，2)＝[A(8，4)+A(8，6)]/2
O(3，3)＝[A(8，7)+A(8，8)]/2
It should be noted that, pixel during pixel in 7th, 8 row of the 1st 8 × 8 dot matrix and the 1st of the 2nd 8 × 8 dot matrix the arranges can be combined as 33 × 3 dot matrix, because the pixel quantity in the 7th, 8 row of the 1st 8 × 8 dot matrix is more, so, from the pixel the 7th, 8 row of the 1st 8 × 8 dot matrix, select pixel to participate in the computing of output pixel point.In like manner, pixel in 1st row of the 7th of the 1st 8 × 8 dot matrix the, the pixel in 8 row and the 4th 8 × 8 dot matrix can be combined as 33 × 3 dot matrix, due to the 1st 8 × 8 dot matrix the 7th, pixel quantity in 8 row is more, so, select pixel to participate in the computing of output pixel point in the pixel from the 7th of the 1st 8 × 8 dot matrix, 8 row.
In this embodiment, the basic templates of the 2nd 8 × 8 dot matrix as shown in Figure 6B.Because the pixel in the 7th, 8 row of the 1st 8 × 8 dot matrix has exported 3 pixels, so the pixel in the 1st row of the 2nd 8 × 8 dot matrix does not participate in the computing of output pixel point in the 2nd 8 × 8 dot matrix.In like manner, pixel during pixel in 8th row of the 2nd 8 × 8 dot matrix and the 1st, 2 of the 3rd 8 × 8 dot matrix arranges also can be combined as 33 × 3 dot matrix, because the pixel quantity in the 1st, 2 row of the 3rd 8 × 8 dot matrix is more, so, pixel in 1st, 2 row of the 3rd 8 × 8 dot matrix exports 3 pixels, and the pixel in the 8th row of the 2nd 8 × 8 dot matrix does not participate in the computing of output pixel point in the 2nd 8 × 8 dot matrix.
Therefore, the 2nd 8 × 8 dot matrix output 2 × 3 pixels, 2 × 3 pixels are as follows respectively:
O(1，4)＝[A(2，2)+A(2，4)]/2
O(1，5)＝[A(2，5)+A(2，7)]/2
O(2，4)＝[A(5，2)+A(5，4)]/2
O(2，5)＝[A(5，5)+A(5，7)]/2
O(3，4)＝[A(8，2)+A(8，4)]/2
O(1，5)＝[A(8，5)+A(8，7)]/2
In this embodiment, the basic templates of the 4th 8 × 8 dot matrix is respectively as shown in Fig. 6 B and Fig. 6 C.Due to the 7th of the 1st 8 × 8 dot matrix the, the pixel in 8 row exported 3 pixels, so the pixel in the 1st row of the 4th 8 × 8 dot matrix does not participate in the computing of output pixel point in this 8 × 8 dot matrix.In like manner, the 1st of pixel in the eighth row of this 8 × 8 dot matrix and the 7th 8 × 8 dot matrix, the pixel in 2 row also can be combined as 33 × 3 dot matrix, due to the 7th 8 × 8 dot matrix the 7th, pixel quantity in 8 row is more, so, the 7th of 7th 8 × 8 dot matrix, the pixel in 8 row exports 3 pixels, and the pixel in the eighth row of the 4th 8 × 8 dot matrix does not participate in the computing of output pixel point in this 8 × 8 dot matrix.
Therefore, the 4th 8 × 8 dot matrix output 3 × 2 pixels, the value of 3 × 2 pixels is as follows respectively:
O(4，1)＝[A(3，1)+A(3，3)]/2
O(4，2)＝[A(3，4)+A(3，6)]/2
O(4，3)＝[A(3，7)+A(3，8)]/2
O(5，1)＝[A(6，1)+A(6，3)]/2
O(5，2)＝[A(6，4)+A(6，6)]/2
O(5，3)＝[A(6，7)+A(6，8)]/2
The value of the pixel of other 8 × 8 dot matrix outputs and above-mentioned computational methods in like manner, repeat no more here.
Again for N=5, medium processing device needs the picture signal received to reduce 5 times.In this case, 8 × 8 dot matrix need output 1.6 × 1.6 pixels, but in actual applications, do not have the pixel of non-integer number.To this, the least common multiple getting 5 and 8 is 40, like this, and 5 × 58 × 8 dot matrix outputs, 8 × 8 pixels.In this embodiment, 5 × 58 × 8 dot matrix can be divided into 8 × 85 × 5 dot matrix, and each 5 × 5 dot matrix export a pixel.The computing of the value of the pixel of each 5 × 5 dot matrix outputs with reference to disposition during N=3, can repeat no more here.
If N is the multiple of 8, so k × k 8 × 8 dot matrix composition N × N dot matrix, wherein, k=N/8.In this case, medium processing device can obtain intermediate object program, and intermediate object program is corresponding with N × N dot matrix that 8 × 8 dot matrix of current acquisition participate in forming.
If N is the multiple of 8, after so medium processing device obtains the intermediate object program of corresponding N × N dot matrix, intermediate object program can be stored.Medium processing device, after all intermediate object programs obtaining corresponding N × N dot matrix, can carry out computing to all intermediate object program, obtains the pixel of corresponding N × N dot matrix.Such as, medium processing device can carry out summation operation to the pixel chosen from current 8 × 8 dot matrix obtained, the result of summation operation is intermediate object program, medium processing device can carry out mean operation to the result of all summation operation, obtain the pixel of corresponding N × N dot matrix, wherein, the denominator that mean operation uses is the number of the pixel of all participation computings, that is, 8 × 8 dot matrix of current acquisition participate in the number of the pixel of all participation computings in N × N dot matrix of composition.
For N=16, medium processing device needs the picture signal received to reduce 16 times.In this case, 8 × 8 dot matrix need output 0.5 × 0.5 pixel, but in actual applications, do not have the pixel of non-integer number.To this, the least common multiple getting 16 and 8 is 16, and like this, 2 × 28 × 8 dot matrix outputs, 2 × 2 pixels, concrete division as shown in Figure 7.In this embodiment, each 2 × 28 × 8 dot matrix export a pixel.Those skilled in the art according to actual needs, can design the pixel of the computing participating in output pixel point, use mean value compute mode, obtain the value of the pixel exported.It should be noted that, the order of the neighbouring relations between the row belonging to 8 × 8 dot matrix in Fig. 7,8 × 8 dot matrix and reception 8 × 8 dot matrix see the associated description in the embodiment of above-mentioned N=3, can repeat no more here.
It should be noted that, because medium processing device receives 8 × 8 dot matrix of 4 shown in Fig. 7 respectively according to sequencing, so, if all there is pixel to need to participate in the computing of output pixel point in 48 × 8 dot matrix, so medium processing device is after acquisition the 1st 8 × 8 dot matrix, the value participating in the pixel of the computing of output pixel point in 1st 8 × 8 dot matrix can be added, buffered results, after acquisition the 2nd 8 × 8 dot matrix, the value participating in the pixel of the computing of output pixel point in 2nd 8 × 8 dot matrix can be added, buffered results, after acquisition the 3rd 8 × 8 dot matrix, the value participating in the pixel of the computing of output pixel point in 3rd 8 × 8 dot matrix can be added, buffered results, after acquisition the 4th 8 × 8 dot matrix, the value participating in the pixel of the computing of output pixel point in 4th 8 × 8 dot matrix can be added, obtain a result, finally, above-mentioned 4 results are done mean value computing, obtain the value of the pixel exported, wherein, the denominator that mean value computing uses is the 1st 8 × 8 dot matrix, 2nd 8 × 8 dot matrix, the number of the pixel of all participation computings in 3rd 8 × 8 dot matrix and the 4th 8 × 8 dot matrix.
If N be greater than 8 and be not 8 multiple, so N × N dot matrix is made up of the boundary pixel point at least one 8 × 8 dot matrix and 8 × 8 adjacent dot matrix, wherein, pixel here refers to that the pixel in 8 × 8 dot matrix in 8 × 8 dot matrix with adjacent together participates in forming the pixel of N × N dot matrix.
If all pixels in 8 × 8 dot matrix of current acquisition all participate in composition N × N dot matrix, so medium processing device can from all pixels current 8 × 8 dot matrix obtained selected pixels point, and computing is carried out to the pixel chosen from all pixels in current 8 × 8 dot matrix obtained, obtain the intermediate object program corresponding with N × N dot matrix that 8 × 8 dot matrix of current acquisition participate in forming.
If the boundary pixel point in 8 × 8 dot matrix of current acquisition participates in composition N × N dot matrix, then medium processing device selected pixels point from the boundary pixel point current 8 × 8 dot matrix obtained, and computing is carried out to the pixel chosen from the boundary pixel point in current 8 × 8 dot matrix obtained, obtain the intermediate object program corresponding with N × N dot matrix that 8 × 8 dot matrix of current acquisition participate in forming.
If N be greater than 8 and be not 8 multiple, so medium processing device is after the intermediate object program obtaining corresponding N × N dot matrix, can store intermediate object program.Medium processing device, after all intermediate object programs obtaining corresponding N × N dot matrix, can carry out computing to all intermediate object program, obtains the pixel of corresponding N × N dot matrix.Such as, medium processing device can carry out summation operation to the pixel chosen from current 8 × 8 dot matrix obtained, the result of summation operation is intermediate object program, medium processing device can carry out mean operation to the result of all summation operation, obtain the pixel of corresponding N × N dot matrix, wherein, the denominator that mean operation uses is the number of the pixel of all participation computings, that is, 8 × 8 dot matrix of current acquisition participate in the number of the pixel of all participation computings in N × N dot matrix of composition.
For N=12, medium processing device needs the picture signal received to reduce 12 times, and in this case, 8 × 8 dot matrix need output 0.67 × 0.67 pixel, but in actual applications, do not have the pixel of non-integer number.To this, the least common multiple getting 12 and 8 is 24, like this, and 3 × 38 × 8 dot matrix outputs, 2 × 2 pixels.In this embodiment, 3 × 38 × 8 dot matrix can be divided into 2 × 2 12 × 12 dot matrix, and each 12 × 12 dot matrix export a pixel.As shown in Figure 8, pixel in the 1-4 row that the 1st 12 × 12 dot matrix after division are arranged by the 1-4 of the 1st 8 × 8 dot matrix, the 2nd 8 × 8 dot matrix, the 1-4 1-4 that is capable and the 5th 8 × 8 dot matrix of the 4th 8 × 8 dot matrix is capable forms, and the theory of constitution of other 12 × 12 dot matrix is identical therewith.Those skilled in the art according to actual needs, can design the pixel of the computing of participation the 1st 12 × 12 dot matrix output pixels, use mean value compute mode, obtain the value of the pixel of the 1st 12 × 12 dot matrix outputs.In like manner, the value of the pixel of other 12 × 12 dot matrix outputs also can be obtained by computing.It should be noted that, the order of the neighbouring relations between the row belonging to 8 × 8 dot matrix in Fig. 8,8 × 8 dot matrix and reception 8 × 8 dot matrix see the associated description in the embodiment of above-mentioned N=3, can repeat no more here.
It should be noted that, because medium processing device receives 8 × 8 dot matrix shown in Fig. 8 respectively according to sequencing, so, if the 1st 8 × 8 dot matrix, the 1-4 row of the 2nd 8 × 8 dot matrix, pixel is all had to need to participate in the computing of output pixel point in the 1-4 row that the 1-4 of the 4th 8 × 8 dot matrix 1-4 that is capable and the 5th 8 × 8 dot matrix is capable, so medium processing device is after acquisition the 1st 8 × 8 dot matrix, the value participating in the pixel of the computing of output pixel point in 1st 8 × 8 dot matrix can be added, buffered results, after acquisition the 2nd 8 × 8 dot matrix, the value of the pixel participating in the computing of the 1st 12 × 12 dot matrix output pixels in the 2nd 8 × 8 dot matrix can be added, buffered results, after acquisition the 4th 8 × 8 dot matrix, the value of the pixel participating in the computing of the 1st 12 × 12 dot matrix output pixels in the 4th 8 × 8 dot matrix can be added, buffered results, after acquisition the 5th 8 × 8 dot matrix, the value of the pixel participating in the computing of the 1st 12 × 12 dot matrix output pixels in the 5th 8 × 8 dot matrix can be added, obtain a result, finally, above-mentioned 4 results are done mean value computing, obtain the value of the pixel of the 1st 12 × 12 dot matrix outputs, wherein, the denominator that mean value computing uses is the 1st 8 × 8 dot matrix, 2nd 8 × 8 dot matrix, the number of the pixel of all participation computings in 4th 8 × 8 dot matrix and the 5th 8 × 8 dot matrix.The processing mode of the value of other 12 × 12 dot matrix output pixels is identical therewith, repeats no more here.
It should be noted that, in the above-described embodiments, if N be greater than 8 and be not 8 multiple, so each 8 × 8 dot matrix relating to composition N × N dot matrix are needed to calculate intermediate object program, and all intermediate object programs corresponding to this N × N dot matrix carry out computing, obtain the pixel of this N × N dot matrix corresponding.In actual applications, medium processing device also can obtain the pixel of this N × N dot matrix corresponding by other means.Such as, medium processing device can only selected pixels point from the pixel quantity participating in composition N × N dot matrix maximum 8 × 8 dot matrix, and the pixel chosen is calculated, thus obtain the pixel of corresponding N × N dot matrix, briefly, can think that maximum 8 × 8 dot matrix of the pixel quantity that participates in composition N × N dot matrix represent N × N dot matrix.Or for N=12, as shown in Figure 8, the 1st 12 × 12 dot matrix after division are by the 1st 8 × 8 dot matrix, the 1-4 row of the 2nd 8 × 8 dot matrix, pixel composition in the 1-4 row that the 1-4 of the 4th 8 × 8 dot matrix 1-4 that is capable and the 5th 8 × 8 dot matrix is capable, when calculating the pixel of corresponding 12 × 12 dot matrix, because the 1st 8 × 8 dot matrix participate in 8 × 8 maximum dot matrix of composition the 1st 12 × 12 lattice pixels point quantity, so, can only selected pixels point from the 1st 8 × 8 dot matrix, and computing is carried out to the pixel chosen, thus obtain the pixel of corresponding 1st 12 × 12 dot matrix.
For making those skilled in the art understanding the present invention clearly, below again composition graphs 9 the present invention will be described.As shown in Figure 9, realization flow of the present invention can comprise:
S901: obtain minification N;
S902: judge whether N is less than 8, if so, then turn S903, otherwise, turn S908;
S903: judge that whether N is the approximate number of 8, if so, then turn S904, otherwise, turn S906;
S904: the basic templates selecting corresponding N × N dot matrix, turns S905;
S905: according to the basic templates selected, selected pixels point from each N × N dot matrix 8 × 8 dot matrix respectively, and respectively computing is carried out to the pixel chosen, obtain the pixel corresponding with each N × N dot matrix in 8 × 8 dot matrix respectively, turn S913;
S906: select the basic templates of corresponding N × N dot matrix and the basic templates of corresponding boundary pixel point, turn S907;
S907: according to the basic templates selected, selected pixels point from each N × N dot matrix 8 × 8 dot matrix and boundary pixel point respectively, and respectively computing is carried out to the pixel selected, obtain the pixel corresponding with each N × N dot matrix in 8 × 8 dot matrix and boundary pixel point respectively, turn S913;
S908: judge that whether N is the multiple of 8, if so, then turn S909, otherwise, turn S911;
S909: the basic templates selecting corresponding 8 × 8 dot matrix, turns S910;
S910: according to the basic templates selected, respectively from selected pixels point each 8 × 8 dot matrix participating in composition N × N dot matrix, and respectively computing is carried out to the pixel chosen, obtain multiple intermediate object program, wherein, participate in the corresponding intermediate object program of each 8 × 8 dot matrix of composition N × N dot matrix, afterwards, computing is carried out to all intermediate object program, obtains the pixel of corresponding N × N dot matrix, turn S913;
S911: the basic templates selecting all pixels in corresponding 8 × 8 dot matrix participating in composition N × N dot matrix and the boundary pixel point in 8 × 8 dot matrix, turns S912;
S912: according to the basic templates selected, respectively from selected pixels point each 8 × 8 dot matrix participating in composition N × N dot matrix, and respectively computing is carried out to the pixel chosen, obtain multiple intermediate object program, wherein, participate in the corresponding intermediate object program of each 8 × 8 dot matrix of composition N × N dot matrix, afterwards, computing is carried out to all intermediate object program, obtains the pixel of corresponding N × N dot matrix, turn S913;
S913: export the pixel obtained.
Corresponding to the embodiment of the method shown in Fig. 3, the present invention also provides a kind of medium processing device.As shown in Figure 10, this medium processing device comprises: obtain unit 1001, is the picture signal of unit for obtaining with 8 × 8 dot matrix; Choose unit 1002, for selected pixels point from current 8 × 8 dot matrix obtained, the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in by the pixel chosen, and wherein, N is positive integer; Arithmetic element 1003, for carrying out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix; Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.Mentioned above, medium processing device is that unit receives image with 8 × 8 dot matrix, so what obtain that unit 1001 obtains is 8 × 8 dot matrix of composition diagram picture.After medium processing device obtains 8 × 8 dot matrix, can buffer memory 8 × 8 dot matrix, so, medium processing device can also comprise the buffer unit (not illustrating in Figure 10) that has caching function, after acquisition unit 1001 obtains 8 × 8 dot matrix, 8 × 8 dot matrix can be issued buffer unit and carry out buffer memory.Choose unit 1002 and can choose the pixel participating in computing from 8 × 8 dot matrix of buffer unit buffer memory.
Above-mentioned medium processing device can also comprise memory cell 1004, for storing the intermediate object program that arithmetic element 1003 obtains.Arithmetic element 1003, after all intermediate object programs obtaining composition N × N dot matrix, can be carried out computing to all intermediate object program, obtain the pixel of corresponding N × N dot matrix.
Arithmetic element 1003 specifically may be used for carrying out mean operation to choosing the pixel that unit 1002 chooses.
Arithmetic element 1003 specifically may be used for carrying out summation operation to choosing unit 1002 from obtaining the pixel chosen current 8 × 8 dot matrix obtained of unit 1001.The intermediate object program of corresponding N × N dot matrix is arithmetic element 1003 to choosing unit 1002 from the result obtaining the pixel chosen current 8 × 8 dot matrix obtained of unit 1001 and carry out summation operation.Arithmetic element 1003 specifically may be used for carrying out mean operation to the result of all summation operation corresponding with N × N dot matrix, obtains the pixel of corresponding N × N dot matrix, and wherein, the denominator that mean operation uses is the number of the pixel of all participation computings.
Because the embodiment of above-mentioned medium processing device is corresponding with the embodiment of the method shown in Fig. 3, so, the function of the unit in above-mentioned medium processing device and the relation of cooperatively interacting can be shown in Figure 3 embodiment of the method in associated description, repeat no more here.
In addition, the present invention also provides a kind of image processing system, comprises decoder, display device and above-mentioned medium processing device.Matching relationship between the function of medium processing device and medium processing device and decoder, display device see description above, can repeat no more here.
In sum, in the present invention, after medium processing device obtains and is the picture signal of unit with 8 × 8 dot matrix, the pixel for participating in the picture signal taking 8 × 8 dot matrix as unit to reduce N process is doubly chosen from current 8 × 8 dot matrix obtained, computing is carried out to the pixel chosen, exports pixel or the intermediate object program of corresponding N × N dot matrix, need not after multiple 8 × 8 dot matrix of buffer memory, carry out again reducing process, save the cache resources needed for multiple 8 × 8 dot matrix of buffer memory.
In addition, in the present invention, medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained, and obtains pixel or intermediate object program, does not need to carry out anti-aliasing filter, realizes comparatively simple and direct.
Through the above description of the embodiments, those skilled in the art can be well understood to the present invention can pass through hardware implementing, the mode that also can add necessary general hardware platform by software realizes, based on such understanding, technical scheme of the present invention can embody with the form of software product, it (can be CD-ROM that this software product can be stored in a non-volatile memory medium, USB flash disk, portable hard drive etc.) in, comprising some instructions in order to make a computer equipment (can be personal computer, server, or the network equipment etc.) perform method described in the present invention each embodiment.
In a word, the foregoing is only preferred embodiment of the present invention, be not intended to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (17)
1. an image downscaling method, it is characterized in that, be applicable to image preview mode, be applied to the image processing system comprising decoder and medium processing device, what medium processing device exported decoder is unit with 8 × 8 dot matrix picture signal reduces process, the described picture signal being unit with 8 × 8 dot matrix is the picture signal that decoder obtains after discrete cosine transform decoding, and described method comprises:
It is the picture signal of unit that medium processing device obtains with 8 × 8 dot matrix;
Medium processing device is selected pixels point from current 8 × 8 dot matrix obtained, and the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in, wherein, N is positive integer;
Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained;
If N is less than 8, then obtain the pixel of corresponding N × N dot matrix;
If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix;
Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.
2. the method for claim 1, is characterized in that, if N is the approximate number of 8, then 8 × 8 dot matrix of current acquisition are divided into the individual N × N dot matrix of (8/N) × (8/N).
3. the method for claim 1, it is characterized in that, if N be less than 8 and be not 8 approximate number, then 8 × 8 dot matrix of current acquisition are divided at least one N × N dot matrix and multiple boundary pixel point, and the boundary pixel point in the boundary pixel point in 8 × 8 dot matrix of current acquisition and 8 × 8 adjacent dot matrix forms N × N dot matrix.
4. method as claimed in claim 2 or claim 3, it is characterized in that, medium processing device selected pixels point from current 8 × 8 dot matrix obtained is specially: medium processing device is selected pixels point from each N × N dot matrix current 8 × 8 dot matrix obtained respectively;
Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained and is specially: medium processing device carries out computing to the pixel chosen from each N × N dot matrix in current 8 × 8 dot matrix obtained respectively;
The pixel that medium processing device obtains corresponding N × N dot matrix is specially: medium processing device obtains the pixel of each N × N dot matrix in 8 × 8 dot matrix of corresponding current acquisition respectively.
5. method as claimed in claim 3, it is characterized in that, medium processing device is selected pixels point from current 8 × 8 dot matrix obtained in the following manner: if many for the quantity for the boundary pixel point forming described N × N dot matrix in adjacent 8 × 8 dot matrix of the number ratio of the boundary pixel point forming N × N dot matrix in 8 × 8 dot matrix of current acquisition, then medium processing device from current 8 × 8 dot matrix obtained for form described N × N dot matrix boundary pixel point in selected pixels point;
Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained and is specially: medium processing device carries out computing to from current 8 × 8 dot matrix obtained for forming the pixel chosen in the boundary pixel point of described N × N dot matrix;
Medium processing device obtains the pixel of corresponding described N × N dot matrix.
6. the method for claim 1, is characterized in that, if N is the multiple of 8, then and k × k 8 × 8 dot matrix composition N × N dot matrix, wherein, k=N/8.
7. method as claimed in claim 6, it is characterized in that, the intermediate object program that medium processing device obtains corresponding N × N dot matrix is specially: medium processing device obtains intermediate object program, and described intermediate object program is corresponding with N × N dot matrix that 8 × 8 dot matrix of current acquisition participate in forming.
8. the method for claim 1, is characterized in that, if N be greater than 8 and be not 8 multiple, then N × N dot matrix is made up of the boundary pixel point at least one 8 × 8 dot matrix and 8 × 8 adjacent dot matrix.
9. method as claimed in claim 8, it is characterized in that, medium processing device selected pixels point from current 8 × 8 dot matrix obtained is specially: if all pixels in 8 × 8 dot matrix of current acquisition all participate in composition N × N dot matrix, then medium processing device selected pixels point from all pixels current 8 × 8 dot matrix obtained; If the boundary pixel point in 8 × 8 dot matrix of current acquisition participates in composition N × N dot matrix, then medium processing device selected pixels point from the boundary pixel point current 8 × 8 dot matrix obtained;
Medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained and is specially: if all pixels in 8 × 8 dot matrix of current acquisition all participate in composition N × N dot matrix, then medium processing device carries out computing to the pixel chosen from all pixels in current 8 × 8 dot matrix obtained; If the boundary pixel point in 8 × 8 dot matrix of current acquisition participates in composition N × N dot matrix, then medium processing device carries out computing to the pixel chosen from the boundary pixel point in current 8 × 8 dot matrix obtained;
The intermediate object program that medium processing device obtains corresponding N × N dot matrix is specially: medium processing device obtains the intermediate object program corresponding with N × N dot matrix that 8 × 8 dot matrix of current acquisition participate in forming.
10. the method as described in claim 1,6-9 any one, is characterized in that, medium processing device stores described intermediate object program after obtaining the intermediate object program of corresponding N × N dot matrix;
Medium processing device, after all intermediate object programs obtaining corresponding N × N dot matrix, carries out computing to all intermediate object program, obtains the pixel of corresponding N × N dot matrix.
11. methods as described in claim 1-3,5-9 any one, it is characterized in that, medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained and is specially: medium processing device carries out mean operation to the pixel chosen.
12. methods as claimed in claim 10, it is characterized in that, medium processing device carries out computing to the pixel chosen from current 8 × 8 dot matrix obtained and is specially: medium processing device carries out summation operation to the pixel chosen from current 8 × 8 dot matrix obtained;
The intermediate object program of described corresponding N × N dot matrix is that medium processing device carries out the result of summation operation to the pixel chosen from current 8 × 8 dot matrix obtained;
Medium processing device carries out computing to all intermediate object program, obtain the pixel of corresponding N × N dot matrix is specially: medium processing device carries out mean operation to the result of all summation operation, obtain the pixel of corresponding N × N dot matrix, wherein, the denominator that mean operation uses is the number of the pixel of all participation computings.
13. 1 kinds of medium processing devices, it is characterized in that, be applicable to image preview mode, be applied to the image processing system comprising decoder and medium processing device, what medium processing device exported decoder is unit with 8 × 8 dot matrix picture signal reduces process, the described picture signal being unit with 8 × 8 dot matrix is the picture signal that decoder obtains after discrete cosine transform decoding, and described medium processing device comprises:
Obtaining unit, is the picture signal of unit for obtaining with 8 × 8 dot matrix;
Choose unit, for selected pixels point from current 8 × 8 dot matrix obtained, the picture signal taking 8 × 8 dot matrix as unit is reduced N process doubly for participating in by the pixel chosen, and wherein, N is positive integer;
Arithmetic element, for carrying out computing to the pixel chosen from current 8 × 8 dot matrix obtained; If N is less than 8, then obtain the pixel of corresponding N × N dot matrix; If N is greater than 8, then obtain the intermediate object program of corresponding N × N dot matrix;
Wherein, obtain a pixel for each N × N dot matrix, if N is greater than 8, then the pixel obtained for each N × N dot matrix carries out computing by multiple intermediate object program and obtains.
14. medium processing devices as claimed in claim 13, is characterized in that, also comprise: memory cell, for storing the intermediate object program that described arithmetic element obtains;
Described arithmetic element, after all intermediate object programs obtaining composition N × N dot matrix, is carried out computing to all intermediate object program, is obtained the pixel of corresponding N × N dot matrix.
15. medium processing devices as claimed in claim 13, is characterized in that, described arithmetic element is specifically for carrying out mean operation to the described pixel choosing unit selection.
16. medium processing devices as claimed in claim 14, is characterized in that, described arithmetic element is specifically for carrying out summation operation to described unit of choosing from the described pixel chosen current 8 × 8 dot matrix obtained of unit that obtains;
The intermediate object program of described corresponding N × N dot matrix is that described arithmetic element obtains from described the result that the pixel chosen current 8 × 8 dot matrix obtained of unit carries out summation operation to described unit of choosing;
Described arithmetic element, specifically for carrying out mean operation to the result of all summation operation corresponding with N × N dot matrix, obtains the pixel of corresponding N × N dot matrix, and wherein, the denominator that mean operation uses is the number of the pixel of all participation computings.
17. 1 kinds of image processing systems, is characterized in that, comprise decoder, display device and the medium processing device described in claim 13-16 any one.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2010/072270 WO2011134144A1 (en) | 2010-04-28 | 2010-04-28 | Image downscaling method and apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102893592A CN102893592A (en) | 2013-01-23 |
| CN102893592B true CN102893592B (en) | 2015-09-30 |
Family
ID=44860757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201080061138.2A Active CN102893592B (en) | 2010-04-28 | 2010-04-28 | Image downscaling method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN102893592B (en) |
| WO (1) | WO2011134144A1 (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1414793A (en) * | 2001-10-23 | 2003-04-30 | 三星电子株式会社 | Compression video decoder with contraction image function and its method |
| US7050656B2 (en) * | 2002-12-11 | 2006-05-23 | Seiko Epson Corporation | Image scaling in the compressed domain |
| CN101483036A (en) * | 2008-01-10 | 2009-07-15 | 扬智科技股份有限公司 | Apparatus for image reduction and method thereof |
| CN101692336A (en) * | 2009-10-09 | 2010-04-07 | 青岛海信宽带多媒体技术有限公司 | Method for zooming out images |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050175258A1 (en) * | 2004-02-11 | 2005-08-11 | Leung Patrick W. | Method and apparatus for downscaling digital image data to fit a graphics display device |
-
2010
- 2010-04-28 CN CN201080061138.2A patent/CN102893592B/en active Active
- 2010-04-28 WO PCT/CN2010/072270 patent/WO2011134144A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1414793A (en) * | 2001-10-23 | 2003-04-30 | 三星电子株式会社 | Compression video decoder with contraction image function and its method |
| US7050656B2 (en) * | 2002-12-11 | 2006-05-23 | Seiko Epson Corporation | Image scaling in the compressed domain |
| CN101483036A (en) * | 2008-01-10 | 2009-07-15 | 扬智科技股份有限公司 | Apparatus for image reduction and method thereof |
| CN101692336A (en) * | 2009-10-09 | 2010-04-07 | 青岛海信宽带多媒体技术有限公司 | Method for zooming out images |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102893592A (en) | 2013-01-23 |
| WO2011134144A1 (en) | 2011-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kim et al. | A real-time convolutional neural network for super-resolution on FPGA with applications to 4K UHD 60 fps video services | |
| CN101123723B (en) | Digital video decoding method based on image processor | |
| CN100412907C (en) | Low complexity and unified transforms for video coding | |
| CN102387360B (en) | Video coding-decoding inter-frame image prediction method and Video Codec | |
| CN103297767A (en) | JPEG image decoding method and decoder suitable for multi-core embedded type platform | |
| CN110996171B (en) | Training data generation method and device for video tasks and server | |
| CN108235057A (en) | A kind of network transmission and method for visualizing of extensive utilizing ocean current data | |
| CN113556442B (en) | Video denoising method and device, electronic equipment and computer readable storage medium | |
| CN104717509A (en) | Method and device for decoding video | |
| CN112702592A (en) | End-to-end binocular image joint compression method, device, equipment and medium | |
| CN103929648A (en) | Motion estimation method and device in frame rate up conversion | |
| CN105208394B (en) | A kind of real-time digital image compression prediction technique and system | |
| CN1964490A (en) | A filter and filtering method | |
| CN106303544A (en) | A kind of video coding-decoding method, encoder | |
| CN102687175B (en) | Decoding system and method operable on encoded texture element blocks | |
| CN102893592B (en) | Image downscaling method and device | |
| CN101256668B (en) | Method for implementing video filtering to working balanced multiple nucleus | |
| CN108288247A (en) | A kind of realtime graphic detail enhancing method on multi-core DSP platform | |
| CN104952038A (en) | SSE2 (streaming SIMD extensions 2nd) instruction set based image interpolation method | |
| CN116257208A (en) | Method and apparatus for separable convolution filter operation on matrix multiplication array | |
| CN101500168B (en) | 1/3 interpolation apparatus for image brightness | |
| CN101370142A (en) | Image decoding and zooming method and device, image display system | |
| CN101389030B (en) | Encoding method, system and video encoder | |
| CN115546036A (en) | Image enhancement method, device, equipment and computer readable storage medium | |
| US8285774B2 (en) | Operation method and apparatus for performing overlap filter and core transform |
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 | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 4, building 406, room 518057, two software park, hi tech Zone, Nanshan District, Guangdong, Shenzhen Patentee after: Shenzhen Shenyang electronic Limited by Share Ltd Address before: 4, building 406, room 518057, two software park, hi tech Zone, Nanshan District, Guangdong, Shenzhen Patentee before: Aike Chuangxin Microelectronic Co., Ltd. |