CN100578544C - Method for sampling the digital image with 9:4 scale - Google Patents
Method for sampling the digital image with 9:4 scale Download PDFInfo
- Publication number
- CN100578544C CN100578544C CN200810020027A CN200810020027A CN100578544C CN 100578544 C CN100578544 C CN 100578544C CN 200810020027 A CN200810020027 A CN 200810020027A CN 200810020027 A CN200810020027 A CN 200810020027A CN 100578544 C CN100578544 C CN 100578544C
- Authority
- CN
- China
- Prior art keywords
- image
- constitutes
- arithmetic mean
- sampling
- image block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention relates to a method for carrying out 9 to 4 down-sapling on a digital image which includes the following steps: A. an inputted image is divided into a plurality of image blocks with size of 3 multiplied by 3; B. 9 to 4 nonlinearity downsapling is carried out on the image obtained by step a to obtain image blocks with size of 2 multiplied by 2; C. the image blocks obtained by step B are arranged according to the positions of the corresponding image blocks before downsapling and recombining a complete output image. The invention aims at providing a method for carrying out 9 to 4 downsapling on the digital image which has small calculation capacity, can not loss the edge of a target and can effectively maintain the details in the image so as to lead the content of the image to adapt the requirements of a display with smaller resolution factor and other applications.
Description
Technical field
The present invention relates to a kind of disposal route of digital picture, be specifically related to a kind of method of digital picture down-sampling.
Background technology
During video and rest image were handled, down-sampling and up-sampling technology were widely used in before the coding step or after the decoding step, so that make the content of image be fit to the spatial resolution of display or the needs of other application.At present, the most frequently used Downsapling method is neighbor interpolation method and bilinear interpolation.
The thought of neighbor interpolation method is: in calculating output image during the pixel value of specified point, the coordinate of this point is obtained a floating-point coordinate by reciprocal transformation, it is simply rounded, obtain an integer type coordinate, the pixel value of this integer type coordinate corresponding pixel in input picture is exactly the result who needs.When carrying out 9 to 4 down-samplings, its essence is exactly directly to extract the result of 4 pixels after as down-sampling according to certain rule in 9 pixels.The thought of bilinear interpolation is: in calculating output image during the pixel value of specified point, the coordinate of this point is obtained a floating-point coordinate by reciprocal transformation, the pixel value of 4 pixels around this floating-point coordinate in the input picture is carried out bilinear interpolation, thereby obtain the result of needs.
The details of above-mentioned two kinds of algorithms is seen 117 pages in " Digital Image Processing " (process plate) book the 8th chapter, the author: Kenneth R.Castleman, publishing house of Tsing-Hua University published in 1998.The advantage of neighbor interpolation method is that operand is little, and shortcoming is the situation that can there be the lose objects edge in the output image that obtains behind the down-sampling, influences the subjective effect of image.The bilinear interpolation advantage be can lose objects the edge, shortcoming is that the details in the image can fog, and operand is bigger.
Summary of the invention
The objective of the invention is to adapt to the needs of the less display of spatial resolution or other application for the content that makes image, it is little to propose a kind of operand, edge that can lose objects, what can effectively keep details in the image carries out 9 to 4 non-linear Downsapling methods to digital picture.
The present invention is achieved by the following technical solutions:
A kind of digital picture is carried out the method for 9 to 4 down-samplings, comprise digital picture is cut apart, further comprising the steps of:
(1) digital picture with input is divided into the image block of several 3*3 sizes;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11Element constitutes, b
12Element is by a
12, a
13The element arithmetic mean constitutes, b
21Element is by a
21, a
31The element arithmetic mean constitutes, b
22Element is by a
22, a
23, a
32, a
33The element arithmetic mean constitutes;
(3) place the position of the image block of the 2*2 image block that step (2) is obtained before by the down-sampling of correspondence, is reassembled into a complete output image, gets final product.
A kind of digital picture is carried out the method for 9 to 4 down-samplings, the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
12The element arithmetic mean constitutes, b
12Element is by a
13Element constitutes, b
21Element is by a
21, a
22, a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
23, a
33The element arithmetic mean constitutes.
A kind of digital picture is carried out the method for 9 to 4 down-samplings, the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
12, a
21, a
22The element arithmetic mean constitutes, b
12Element is by a
13, a
23The element arithmetic mean constitutes, b
21Element is by a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
33Element constitutes.
A kind of digital picture is carried out the method for 9 to 4 down-samplings, the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
21The element arithmetic mean constitutes, b
12Element is by a
12, a
13, a
22, a
23The element arithmetic mean constitutes, b
21Element is by a
31Element constitutes, b
22Element is by a
32, a
33The element arithmetic mean constitutes.
The beneficial effect of the inventive method is, image down sampling can be grown up and wide all be the downscaled images of life size 2/3 size.Make the content of image can adapt to the needs of the less display of spatial resolution or other application.Compare with the operand of known Downsapling method, 4 points of the every computing of this method only need 5 sub-additions and 3 shift operations.1 point of the every computing of bilinear interpolation just needs 4 sub-additions, 4 subtractions and 4 multiplication (to see the formula 6 of 119 pages in " Digital Image Processing " (process plate) book the 8th chapter, author: Kenneth R.Castleman, publishing house of Tsing-Hua University, published in 1998), so this method operand is little a lot.Compare with the effect of known Downsapling method, the edge that this method can lose objects can effectively keep the details of image.
To utilize this that digital picture is carried out the progress of the method and the existing method of 9 to 4 down-samplings in order further proving, to have done following experiment, concrete outcome is referring to accompanying drawing 1~4.
In the accompanying drawing, the word segment that superposes in Fig. 2 image owing to lose the edge, is beyond recognition behind the down-sampling.The right side top of car is watched in Fig. 3 and Fig. 4 contrast, and the details of Fig. 4 is than the clarity of detail of Fig. 3; Die Jia word segment in addition, Fig. 4 and Fig. 3 can recognize, but the details of Fig. 4 is than the clarity of detail of Fig. 3.Therefore this experimental results show that the down-sampling better effects if of this method.
Description of drawings
Fig. 1 carries out the preceding image of digital picture down-sampling.
Fig. 2 is the output image of neighbor interpolation method down-sampling.
Fig. 3 is the output image of bilinear interpolation down-sampling.
Fig. 4 is the output image of the inventive method down-sampling.
Fig. 5 is the process flow diagram of the inventive method.
Embodiment
Referring to shown in Figure 5.
Embodiment 1
Digital picture is carried out the method for 9 to 4 down-samplings, may further comprise the steps:
(1) digital picture with input is divided into the image block of several 3*3 sizes; The method of cutting apart is similar to the piece that in the MPEG4 image compression algorithm image segmentation is become the 16*16 size, and difference only is that the size of this method image block becomes 3*3;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11Element constitutes, b
12Element is by a
12, a
13The element arithmetic mean constitutes, b
21Element is by a
21, a
31The element arithmetic mean constitutes, b
22Element is by a
22, a
23, a
32, a
33The element arithmetic mean constitutes; The physical relationship expression formula is as follows:
[formula 1]
b
11=a
11
[formula 2]
b
12=(a
12+a
13)/2
[formula 3]
b
21=(a
21+a
31)/2
[formula 4]
b
22=(a
22+a
23+a
32+a
33)/4
(3) place the position of the image block of the 2*2 image block that step (2) is obtained before by the down-sampling of correspondence, is reassembled into a complete output image, gets final product.
Embodiment 2
Digital picture is carried out the method for 9 to 4 down-samplings, may further comprise the steps:
(1) digital picture with input is divided into the image block of several 3*3 sizes; The method of cutting apart is similar to the piece that in the MPEG4 image compression algorithm image segmentation is become the 16*16 size, and difference only is that the size of this method image block becomes 3*3;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11, a
12The element arithmetic mean constitutes, b
12Element is by a
13Element constitutes, b
21Element is by a
21, a
22, a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
23, a
33The element arithmetic mean constitutes.The physical relationship expression formula is as follows:
[formula 1]
b
11=(a
11+a
12)/2
[formula 2]
b
12=a
13
[formula 3]
b
21=(a
21+a
22+a
31+a
32)/4
[formula 4]
b
22=(a
23+a
33)/2
(3) place the position of the image block of the 2*2 image block that step (2) is obtained before by the down-sampling of correspondence, is reassembled into a complete output image, gets final product.
Embodiment 3
Digital picture is carried out the method for 9 to 4 down-samplings, may further comprise the steps:
(1) digital picture with input is divided into the image block of several 3*3 sizes; The method of cutting apart is similar to the piece that in the MPEG4 image compression algorithm image segmentation is become the 16*16 size, and difference only is that the size of this method image block becomes 3*3;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11, a
12, a
21, a
22The element arithmetic mean constitutes, b
12Element is by a
13, a
23The element arithmetic mean constitutes, b
21Element is by a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
33Element constitutes.The physical relationship expression formula is as follows:
[formula 1]
b
11=(a
11+a
12+a
21+a
22)/4
[formula 2]
b
12=(a
13+a
23)/2
[formula 3]
b
21=(a
31+a
32)/2
[formula 4]
b
22=a
33
(3) place the position of the image block of the 2*2 image block that step (2) is obtained before by the down-sampling of correspondence, is reassembled into a complete output image, gets final product.
Embodiment 4
Digital picture is carried out the method for 9 to 4 down-samplings, may further comprise the steps:
(1) digital picture with input is divided into the image block of several 3*3 sizes; The method of cutting apart is similar to the piece that in the MPEG4 image compression algorithm image segmentation is become the 16*16 size, and difference only is that the size of this method image block becomes 3*3;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11, a
21The element arithmetic mean constitutes, b
12Element is by a
12, a
13, a
22, a
23The element arithmetic mean constitutes, b
21Element is by a
31Element constitutes, b
22Element is by a
32, a
33The element arithmetic mean constitutes.The physical relationship expression formula is as follows:
[formula 1]
b
11=(a
11+a
21)/2
[formula 2]
b
12=(a
12+a
13+a
22+a
23)/4
[formula 3]
b
21=a
31
[formula 4]
b
22=(a
32+a
33)/2
Claims (4)
1, a kind of digital picture is carried out the method for 9 to 4 down-samplings, comprises digital picture is cut apart, it is characterized in that may further comprise the steps:
(1) digital picture with input is divided into the image block of several 3*3 sizes;
(2) step (1) is cut apart the image block that obtains as the input data, is made as:
Be designated as A; The output data of down-sampling is the image block of 2*2 size, is made as:
Be designated as B; The mode that sampling is chosen is: b
11Element is by a
11Element constitutes, b
12Element is by a
12, a
13The element arithmetic mean constitutes, b
21Element is by a
21, a
31The element arithmetic mean constitutes, b
22Element is by a
22, a
23, a
32, a
33The element arithmetic mean constitutes;
(3) place the position of the image block of the 2*2 image block that step (2) is obtained before by the down-sampling of correspondence, is reassembled into a complete output image, gets final product.
2, according to claim 1ly a kind of digital picture is carried out the method for 9 to 4 down-samplings, it is characterized in that the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
12The element arithmetic mean constitutes, b
12Element is by a
13Element constitutes, b
21Element is by a
21, a
22, a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
23, a
33The element arithmetic mean constitutes.
3, according to claim 1ly a kind of digital picture is carried out the method for 9 to 4 down-samplings, it is characterized in that the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
12, a
21, a
22The element arithmetic mean constitutes, b
12Element is by a
13, a
23The element arithmetic mean constitutes, b
11Element is by a
31, a
32The element arithmetic mean constitutes, b
22Element is by a
33Element constitutes.
4, according to claim 1ly a kind of digital picture is carried out the method for 9 to 4 down-samplings, it is characterized in that the mode that sampling is chosen described in the described step (2) is: b
11Element is by a
11, a
21The element arithmetic mean constitutes, b
12Element is by a
12, a
13, a
22, a
23The element arithmetic mean constitutes, b
21Element is by a
31Element constitutes, b
22Element is by a
32, a
33The element arithmetic mean constitutes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200810020027A CN100578544C (en) | 2008-03-24 | 2008-03-24 | Method for sampling the digital image with 9:4 scale |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200810020027A CN100578544C (en) | 2008-03-24 | 2008-03-24 | Method for sampling the digital image with 9:4 scale |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101261731A CN101261731A (en) | 2008-09-10 |
CN100578544C true CN100578544C (en) | 2010-01-06 |
Family
ID=39962173
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN200810020027A Active CN100578544C (en) | 2008-03-24 | 2008-03-24 | Method for sampling the digital image with 9:4 scale |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100578544C (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB201007580D0 (en) * | 2010-05-06 | 2010-06-23 | Im Sense Ltd | Making robust images |
CN104423783B (en) * | 2013-09-02 | 2019-03-29 | 联想(北京)有限公司 | The method and electronic equipment of information transmission |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4356555A (en) * | 1978-03-24 | 1982-10-26 | Ricoh Co., Ltd. | Method of restoring a picture cell by estimation with high density |
-
2008
- 2008-03-24 CN CN200810020027A patent/CN100578544C/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4356555A (en) * | 1978-03-24 | 1982-10-26 | Ricoh Co., Ltd. | Method of restoring a picture cell by estimation with high density |
Also Published As
Publication number | Publication date |
---|---|
CN101261731A (en) | 2008-09-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109389556B (en) | Multi-scale cavity convolutional neural network super-resolution reconstruction method and device | |
CN109493350B (en) | Portrait segmentation method and device | |
EP1794714B1 (en) | Magnification and pinching of two-dimensional images | |
KR102292559B1 (en) | Monocular image depth estimation method and apparatus, apparatus, program and storage medium | |
EP3839875B1 (en) | Method and apparatus for correcting distorted document image | |
US9628805B2 (en) | Tunable multi-part perceptual image hashing | |
US20170061677A1 (en) | Disparate scaling based image processing device, method of image processing, and electronic system including the same | |
CN103559697A (en) | Scrap paper lengthwise cutting splicing and recovering algorithm based on FFT | |
Sajjad et al. | Multi-kernel based adaptive interpolation for image super-resolution | |
Upneja et al. | Fast computation of Jacobi-Fourier moments for invariant image recognition | |
CN111738952A (en) | Image restoration method and device and electronic equipment | |
CN112997479A (en) | Method, system and computer readable medium for processing images across a phase jump connection | |
WO2018120082A1 (en) | Apparatus, method and computer program product for deep learning | |
CN111968033A (en) | Image scaling processing method and device | |
CN116258850A (en) | Image semantic segmentation method, electronic device and computer readable storage medium | |
CN100578544C (en) | Method for sampling the digital image with 9:4 scale | |
CN113705575B (en) | Image segmentation method, device, equipment and storage medium | |
Zhang et al. | A GPU-accelerated real-time single image de-hazing method using pixel-level optimal de-hazing criterion | |
US7389005B2 (en) | Interpolation of video and audio digital data | |
CN113963009A (en) | Local self-attention image processing method and model based on deformable blocks | |
US20230098437A1 (en) | Reference-Based Super-Resolution for Image and Video Enhancement | |
CN115908531A (en) | Vehicle-mounted distance measuring method and device, vehicle-mounted terminal and readable storage medium | |
CN104700357A (en) | Chinese character image zooming method based on bilinear operator | |
CN115147283A (en) | Image reconstruction method, device, equipment and medium | |
CN113643173A (en) | Watermark removing method, watermark removing device, terminal equipment and readable storage medium |
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 | ||
C56 | Change in the name or address of the patentee | ||
CP01 | Change in the name or title of a patent holder |
Address after: 230088 National University Park, 669 Changjiang West Road, Anhui, Hefei Patentee after: ANHUI CREARO TECHNOLOGY CO., LTD. Address before: 230088 National University Park, 669 Changjiang West Road, Anhui, Hefei Patentee before: Anhui Crearo Technology Co., Ltd. |