CN101371566A - Method and apparatus for defective pixel detection based on the human visual system - Google Patents
Method and apparatus for defective pixel detection based on the human visual system Download PDFInfo
- Publication number
- CN101371566A CN101371566A CNA2007800030768A CN200780003076A CN101371566A CN 101371566 A CN101371566 A CN 101371566A CN A2007800030768 A CNA2007800030768 A CN A2007800030768A CN 200780003076 A CN200780003076 A CN 200780003076A CN 101371566 A CN101371566 A CN 101371566A
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- test
- threshold value
- bad
- 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 invention discloses a bad pixel detection method and a module which provide a quick-test and a full-test for bad pixel detections in an image. The quick-test tests a current pixel to one and only one good neighbor having been previously tested. The quick-test is optimized by exploiting weaknesses in the human visual system especially for red and blue colors. More lenient thresholds can be used for the blue color compared to thresholds for the red and green colors. Moreover, the full-test is constructed and arranged to detect bad pixel clusters in a kernel.
Description
The CROSS-REFERENCE TO RELATED APPLICATIONS case
The application's case advocates to have precedence over that on January 19th, 2006 filed an application and name is called No. the 60/760th, 669, the U.S. Provisional Application case of " a kind of method (A Method for Defective Pixel Detection Based on theHuman Visual System) that is used for based on human visual system's defectiveness pixel detection ".
Technical field
The present invention relates to the image processing principle, and more particularly, relate to the detection and the correction of the pixel in the digital picture, wherein said digital picture has the in fact color-values in former scene not of possibility.These pixels are because dust granule in CDD or the CMOS optical array element or imperfection but common.Exactly, proposed method relate to use based on the human visual system produce can not the perception difference lenient thresholds carry out each pixel and its tightly around neighbour's quick test.The invention still further relates to bad pixel detection module, and relate to and to carry out to be used for the program command of bad pixel detection by processor.
Background technology
The defectiveness image pixel is defined as the pixel that its response obviously is different from its neighbour's value.Dust granule or microlens defective are two common causes of given pixel possibility reporting errors value.
A kind of common methods that is used for bad pixel detection is big a certain percentage of maximum or the fixed threshold in its neighbour's value of response ratio of given pixel, or under the situation than the little a certain percentage of minimum value of its neighbour's value or fixed threshold it is labeled as defectiveness.This cross range request to system storage repeatedly read access, be maximum and the minimum value of repeatedly logic comparison then to calculate each neighbor neighbour.
Need carry out bad pixel detection with the speed more faster than current known method.Carrying out the speed of bad pixel detection can come existing method is optimized by the weakness among the exploitation human visual system.Briefly, compare with the threshold value of green, redness can be used looser threshold value, and compares with redness and green, and blueness can be used even looser threshold value.
In addition, red, green and the blue thresholds value that can further depend upon candidate pixel loosens.
Summary of the invention
Consider that above a target of the present invention provides a kind of method of carrying out bad pixel detection with the speed more faster than current known method.
Another target of the present invention provides a kind of good neighbour's pixel of same color of using and comes pixel is carried out the bad pixel detection scheme of test fast.
Another target of the present invention provides weakness among a kind of human visual system of exploitation with the optimized bad pixel detection scheme of pixel detection speed.
Another target of the present invention provides a kind of bad pixel detection scheme, and it compares the looser red pixel threshold value of employing with the threshold value of green pixel, and compares employing even looser blue pixel threshold value with redness and green pixel.
Another target of the present invention provides a kind of bad pixel detection scheme, and the value that wherein red, green and blue thresholds can further depend upon candidate pixel loosens.
Aforementioned and other targets of the present invention are implemented by bad pixel detection module, described bad pixel detection module comprises quick test bad pixel detection submodule, its can operate with the pixel value of current pixel in the image and only difference and the threshold value between the pixel value of good neighbour's pixel compare, described threshold value is to select to be used for to create in described image the noticeable bad pixel of zero.
Described bad pixel detection module further comprises complete test bad pixel detection submodule, its can operate with current pixel is tested fully with only current pixel not by the situation of quick test under the assessment current pixel whether be bad pixel in the nuclear.
The present invention also provides the bad pixel detection module that comprises quick test bad pixel detection submodule, and described quick test bad pixel detection submodule can be operated so that difference and max-thresholds between good neighbour's pixel value of the current pixel value of a color in the image and described color are compared.Described max-thresholds is based on the human visual system response of described color is created the noticeable bad pixel of zero in described image.Described module also comprises complete test bad pixel detection submodule, and whether it can operate to be assessed noticeable bad pixel by the test fully of nuclear is bad pixel in the described nuclear.
Bad pixel detection module further comprises quick test bad pixel detection submodule, its can operate with current pixel value and one and only good neighbour's pixel of a first Pretesting compare.Fast testing needle to next not test pixel repeat described quick test.
Described bad pixel detection module also comprises complete test bad pixel detection submodule, to test current pixel again when finding that at quick test period current pixel is noticeable bad pixel.
Described bad pixel detection module is looser than the threshold value of green with the threshold setting of some pixel shader of redness.
Bad pixel detection module is looser than red and green threshold value with the threshold setting of each pixel shader of blueness.In addition, with the value increase of pixel value, the threshold value of pixel value becomes looser.
In another aspect, the present invention is intended to can be by the program command of processor execution, described program command can be operated when carrying out to compare by difference between the pixel value of the pixel value of current pixel in image and good neighbour's pixel only and selected threshold value and test current pixel fast, thus in image the noticeable bad pixel of establishment zero.
Described program command when carrying out, can operate with test fully current pixel only come current pixel not by the situation of quick test under the assessment current pixel whether be bad pixel in the nuclear.
More on the one hand in, the present invention is intended to a kind of method that is used for bad pixel detection, it may further comprise the steps: whether by based on the human visual system response of a color in the image is compared the difference between good neighbour's pixel value of the current pixel value of described color and described color and threshold value creating the noticeable bad pixel of zero in image, and to test to assess described noticeable bad pixel fully to noticeable bad pixel be bad pixel in the nuclear.
Description of drawings
Following being described in detail in when reading in conjunction with the accompanying drawings of aforementioned summary and the preferred embodiments of the present invention will be better understood.For graphic extension purpose of the present invention, in diagram, show currently preferred embodiments.Yet, should be appreciated that the present invention is not limited to shown accurate layout.In diagram:
Fig. 1 shows common Bel's mode arrangement;
The pixel colour code of Fig. 2 A, 2B and 2C graphic extension monochrome, from 0 to 255 shade of described monochrome are divided into heat, warm and cold pixel shader classification;
The general block diagram of Fig. 3 graphic extension one bad pixel detection module, described bad pixel detection module is connected with the bad pixel correction module that is shown as diplopia;
The block diagram of Fig. 4 graphic extension graphics processing unit;
Fig. 5 graphic extension is based on the flow chart of human visual system's quick test bad pixel detection method;
Fig. 6 graphic extension is based on the human visual system and at the flow chart of RGB heat, the warm and quick test bad pixel detection method of optimization that cold pixel shader is optimized.
Embodiment
Although the present invention can allow many multi-form embodiment, this specification and accompanying drawing only disclose some form as using example of the present invention.The present invention does not plan to be restricted to described embodiment, and scope of the present invention will be pointed out in the claims of enclosing.
Hereinafter be used for the method that defective pixels detects and the preferred embodiment of module by the application-specific of the image of LCD (LCD) being described according to the present invention is based on the human visual system.Yet, be understood by those skilled in the art that the present invention is applicable to that also other display unit types of using the rgb pixel scheme or other pixel colors are for demonstration.In detail with reference to diagram, wherein run through and use same numbers indication identical element in full now, show the module that is used for bad pixel detection according to the present invention in Fig. 3 and 4, it generally refers to 10.
Fig. 1 shows common Bel's mode arrangement of being used by present CCD or CMOS manufacturer.In this graphic extension, even number line contains alternative redness and green pixel, and odd-numbered line comprises alternative green and blue pixel.Here, " neighbour " of pixel is defined as neighbor, and it is after filtration with the same part in response to chromatogram.For example, if check the red pixel defectiveness whether that is labeled as R33, may or may unsight adjacent red pixel be those red pixels that are labeled as R11, R31, R51, R13, R53, R15, R35, R55 then.Check the blue pixel that is labeled as B44 whether during defectiveness, adjacent blue pixel comprises those blue pixel that are labeled as B22, B42, B62, B24, B64, B26, B46 and B66.At last, when checking green pixel (for example G43), green pixel neighbors is defined as being labeled as those green pixels of G41, G32, G52, G23, G63, G34, G54 and G45.
In described example, green pixel is the twice of blueness or red pixel.Adjacent green pixel comprises those green pixels G32, G52, G34 and G54, and itself and green pixel G43 are at a distance of one (1).Green pixel G32, G52, G34 and G54 are and the direct turning green pixel in succession of green pixel G43.The distance of green pixel G41, G23, G63 and G45 and green pixel G43 is two (2) individual pixels.In this example, there are eight (8) individual red pixel neighbors, eight (8) individual blue neighbors and eight (8) individual green pixel neighbors.
Because the maximum of the neighbor of calculating pixel and minimum value are expensive processes, then bad pixel detection module 10 comprises quick test bad pixel detection submodule 20, it at first compares just processed current or candidate pixel and its checked neighbour's pixel, with check described the two whether in same threshold value, wherein described difference will not discovered by human visual system (HVS) 80 in described threshold value.If then bad pixel detection module 10 moves on to next pixel and carries out quick test bad pixel detection (being sometimes referred to as " method for rapidly testing " thereafter).
For example, if it is 55 that given green pixel has pixel value, and the green neighbour's in the upper left corner (treated and be known as better pixel) value is 65, the pixel that then has value 55 is necessary or good, if perhaps in fact be defective also unimportant, can not be detected by HVS 80 because difference is too little.
Bad pixel detection module 10 further comprises carries out the complete test bad pixel detection method complete test bad pixel detection submodule 30 of (after this being called " test fully ").In exemplary embodiments, if find that just evaluated pixel is bad in the quick test bad pixel detection submodule 20, then bad pixel detection module 10 will be carried out test fully as described below.Yet, if, proofread and correct described bad pixel by the bad pixel correction module 50 that is shown as diplopia among Fig. 3 determining that a pixel is " bad " after the test fully.
Carrying out when testing fully, all of current or candidate pixel and same color are being remained neighbors compare.In exemplary embodiments, the test bad pixel detection submodule always is to use good neighbour's pixel of the left upper that is positioned at current or candidate pixel to come described current or candidate pixel execution test fast fast.On the other hand, test will be tested current or candidate pixel by remaining seven (7) individual neighbors fully.
Referring now to Fig. 5, existing with the quick test bad pixel detection method 100 of more detailed description.Method 100 starts from being set in the acquisition of step S102 place the not test pixel of current pixel.Then step S104 after the step S102 wherein compares current pixel value and adjacent pixel values preferably.Described good neighbour's pixel value is used for the good neighbour's pixel through first Pretesting of same color.Is step S106 after the step S104, wherein the result of determining step S104 (poor) whether greater than/less than maximum/minimum human visual system's threshold value (HVS_TH).If the described "Yes" (its indication current pixel is good) that is defined as then is step S108 after the step S106 then, wherein determine whether to exist any pixel of not testing as yet.If be defined as "Yes" at step S108 place, step S108 loop or return step S102 then.Otherwise if be defined as "No" at step S108 place, then method 100 finishes.
Return step S106, if be defined as "No" (its indication current pixel is noticeable bad pixel), then being step S110 after the step S106 then, wherein is that bad pixel detection is carried out test fully, wherein tests current pixel at remaining seven (7) individual neighbors.
Method for rapidly testing 100 can be by using looser threshold value for red pixel, and optimize for the use of blue Bayer pixel even looser threshold value, because human eye is more insensitive to the variation at these frequency places.For example, test has shown and has been respectively 16,12,30 redness, green, blue thresholds performance better.Human visual system 80 RGB threshold value 40 is used by quick test bad pixel detection submodule 20.
In bad pixel detection module 10 by utilizing the threshold value among heat, warm and cold look-up table 42A, 42B and the 42C to come the optimization process number of times.Heat, warm and cold look-up table 42A, 42B and 42C are hot, warm or cold and comprise the HVS threshold value of redness, blueness and green pixel based on pixel value.The pixel colour code of Fig. 2 A, 2B and 2C graphic extension monochrome, from 0 to 255 shade of described monochrome are divided into heat, warm and cold shadow category respectively.Pixel value (PV) is the coldest pixel value of zero (0) representative.On the other hand, pixel value is the hottest pixel values of 255 representatives.From 0 to 255 pixel color scope may be at for example arbitrary monochromes such as redness, green, blueness, white, black.
In Fig. 2 A, be labeled as 230,240, in Fig. 2 B, be labeled as 130,140 and in Fig. 2 C, be labeled as 30,40 frame corresponding to the pixel value in following table 1,2 and 3 240,230,140,130,40,30.Each frame all comprise pixel value between whole-colored two adjacent shades roughly half.For example, frame 30 comprises from the pixel value of 26-35 or 25-34.On the other hand, the pixel value of frame 40 comprises pixel value 36-45 or 35-44.Yet table 1,2,3 can comprise in the 0-255 scope or the clauses and subclauses of each pixel value in other scopes of pixel shader.As finding out according to following table 1,2 and 3, the HVS threshold value becomes looser with the pixel value magnitude increase.
These HVS threshold values of pixel value in the heat rating are shown in the following table 1.
Table 1
| Pixel value (heat rating) | 190 | 200 | 210 | 220 | 230 | 240 | 250 |
| Red HVS threshold value | 18 | 18 | 20 | 22 | 22 | 22 | 22 |
| Green HVS threshold value | 18 | 18 | 18 | 20 | 20 | 20 | 20 |
| Blue HVS threshold value | 32 | 32 | 32 | 32 | 32 | 34 | 34 |
The HVS threshold value of the pixel value in the warm range is shown in the following table 2.
Table 2
| Pixel value (warm range) | 90 | 10 | 110 | 120 | 130 | 140 | 150 | 160 | 170 |
| Red HVS threshold value | 16 | 16 | 16 | 16 | 16 | 16 | 16 | 16 | 16 |
| Green HVS threshold value | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 12 | 16 |
| Blue HVS threshold value | 32 | 30 | 30 | 30 | 30 | 30 | 30 | 30 | 32 |
The HVS threshold value of the pixel value in the cold scope is shown in the following table 3.
Table 3
| Pixel value (cold scope) | 10 | 20 | 30 | 40 | 50 | 60 | 70 | 80 |
| Red |
25 | 25 | 25 | 25 | 22 | 22 | 20 | 18 |
| Green |
25 | 25 | 20 | 20 | 20 | 18 | 16 | 16 |
| Blue |
35 | 35 | 35 | 35 | 35 | 32 | 32 | 32 |
Described redness, blueness and green HVS threshold value are used for defining the maximum difference that increases and reduce at quick test period.The result of the comparison between current pixel value and the good neighbour's pixel value can create between two pixel values can be in response to the difference that increases or reduce.The maximum pixel difference of red pixel increases and the minimizing of maximum pixel difference is marked as MAX_PIX__DIFF_INC_R and MAX_PIX_DIFF_DEC_R.The maximum pixel difference of blue pixel increases and the minimizing of maximum pixel difference is marked as MAX_PIX_DIFF_INC_B and MAX_PIX_DIFF_DEC_B.The maximum pixel difference of green pixel increases and the minimizing of maximum pixel difference is marked as MAX_PIX_DIFF_INC_G and MAX_PIX_DIFF_DEC_G.At the pixel value 120 as the warm pixel value of red, green and blue pixel, MAX_PIX_DIFF_INC_R is 16, and MAX_PIX_DIFF_INC_G is 12, and MAX_PIX_DIFF_INC_B is 30.Equally, at the pixel value 120 of redness, green and blue pixel, MAX_PIX_DIFF_DEC_R is-16, and MAX_PIX_DIFF_DEC_G is-12, and MAX_PIX_DIFF_DEC_B is-30.
Threshold value in the above-mentioned table 1,2 and 3 defines MAX_PIX_DIFF_INC and MAX_PIX_DIFF_DEC (it is the negative value of threshold value in the table 1,2 and 3).
Referring now to Fig. 6, now will describe quick test bad pixel detection method 200 in detail through optimizing.Method 200 starts from being set as in the acquisition of step S202 place the not test pixel of current pixel.After the step S202 then is step S204, wherein obtains the pixel value of good neighbour's pixel (GNP).In described example, with good neighbour's pixel in the upper left corner as reference point.After the step S204 then is one among step S206A, 206B and the 206C, is respectively redness, blueness or green to determine the GNP pixel.If the GNP pixel is red, then be step 208A after the step S206A then, wherein obtain the red HVS threshold value of the respective pixel value of GNP according to heat, warm and cold look-up table 42A, 42B and 42C.
For example, if GNP is red and has pixel value to be 30, then red HVS threshold value is 25 (seeing Table 3).Then is step S210A after the step S208A, wherein maximum/minimum HVS threshold value (HVS TH) is made as red HVS threshold value according to heat, warm and cold look-up table 42A, 42B and 42C.
If, then be step 208B after the step S206B then, wherein obtain the blue HVS threshold value of respective pixel value according to heat, warm and cold look-up table 42a, 42b and 42c as determining in step S206B place that the GNP pixel is blue.For example, if GNP is blue and has pixel value to be 230, then blue HVS threshold value is 32 (seeing Table 1).Step S208B then is step S210B, wherein according to heat, warm and cold look-up table 42A, 42B and 42C maximum/minimum HVS threshold value (HVS TH) is made as blue HVS threshold value.
If determine as step S206C place, GNP is a green, then is step 208C after the step S206C then, wherein from green HVS threshold value hot, warm and cold look-up table 42a, 42b and 42c acquisition respective pixel value.For example, if GNP is green and has pixel value to be 40, then green HVS threshold value is 20 (seeing Table 3).Then is step S210C after the step S208C, wherein maximum/minimum HVS threshold value (HVS_TH) is made as green HVS threshold value according to heat, warm and cold look-up table 42A, 42B and 42C.
RGB HVS in the option table 1,2 and 3 is so that produce the noticeable false negative value of zero.
After step S210A, 210B and the 210C then is step S212, wherein current pixel value (red, blue or green) is compared with good neighbour's pixel value.Then is step S214 after the step S212, wherein the result (poor) who comes determining step S212 according to one among step S210A, 21GB and the 210C whether greater than/less than maximum/minimum human visual system's threshold value (HVS TH).As apprehensible, comparison step S212 can be moved to after step S204 and begin immediately.
Definite described result whether greater than/during less than HVS TH, described definite assessment result whether less than MAX_PIX_DIFF_INC or described result whether greater than MAX_FIX_DIFF_DEC.If the described "Yes" (it means that current pixel is good maybe can not noticing) that is defined as then is step S216 after the step S214 then, wherein determine whether to exist any not tested as yet pixel.If be defined as "Yes" at step S216 place, then next not test pixel is wherein assessed in step S216 loop or turn back to step S202.Otherwise if be defined as "No" at step S216 place, then method 200 finishes.
Turning back to step S214, if be defined as "No" (it means that current pixel is bad maybe can noticing), then is step S218 after the step S214 then, is that bad pixel detection is carried out test fully by complete test bad pixel detection submodule 30 wherein.
Referring now to Fig. 4, graphics processing unit 60 comprises advanced RISC machine (ARM) 65 or other processing unit, and it is coupled to program command 70 and display 75, for example (and unrestricted) LCD (LCD).ARM 65 is coupled to image source 55, so that the pixel image of the image processing that stands graphics processing unit 60 to be provided.Program command 70 can be operated to serve as bad pixel detection module 10 and bad pixel correction module 50 when being carried out by ARM 65.Therefore, program command 70 can be operated to carry out quick test and to test fully carry out bad pixel detection when carrying out.
In operation, utilizing HVS 80 to minimize with the minimum value that will calculate each pixel neighbour and peaked needs helps the bad pixel correction on advanced RISC machine (ARM) is quickened 2.6 times.
Be understood by those skilled in the art that by module, method and the program command that this paper discloses, the detection of bad pixel is than obviously realizing apace in routine techniques.Can optimize the speed of bad pixel detection of execution to be better than existing method by utilizing the weakness among the human visual system.To be used for the process that defective pixels detects also more economical than conventional process according to of the present invention, and for example, conventional method requires to calculate the maximum and the minimum value of adjacent pixel values.
This paper provides the above stated specification of the embodiment of the invention for graphic extension and illustrative purposes.It is not hope as exhaustive or limit the invention to the precise forms that is disclosed, and can obtain various modifications and changes according to above-mentioned teaching or enforcement according to the present invention.Described embodiment is through selecting and describing to explain principle of the present invention and application in practice thereof, so that the those skilled in the art can utilize the present invention in each embodiment, and various modifications is considered as being suitable for the special-purpose that contained.Scope of the present invention plans to be defined by appended claims and equivalent thereof.
Claims (38)
1. bad pixel detection module, it comprises:
Quick test bad pixel detection submodule, its can operate with the pixel value of the current pixel in the image and only difference and the threshold value between the pixel value of good neighbour's pixel compare, described threshold value is to select to be used for to create in described image the noticeable bad pixel of zero.
2. module as claimed in claim 1, it further comprises complete test bad pixel detection submodule, and described complete test bad pixel detection submodule can operate whether assess described current pixel be bad pixel in the nuclear only described current pixel tested fully under not by the situation of described quick test at described current pixel.
3. module as claimed in claim 2, wherein said quick test bad pixel detection submodule can be operated to stop the test to described current pixel when described difference exceeds the described scope of described threshold value, and described current pixel is noticeable bad pixel; And described complete test bad pixel detection submodule can be operated to restart the test to described noticeable bad pixel.
4. module as claimed in claim 1, wherein said good neighbour's pixel are meant previous good neighbour's pixel after tested.
5. module as claimed in claim 1, wherein said quick test bad pixel detection submodule further can be operated to stop the test to described current pixel under the situation of the scope that is positioned at described threshold value in described difference, and further can operate to begin the test to new pixel, described new pixel becomes described current pixel.
6. module as claimed in claim 1, it further comprises at least one look-up table that is used to select described threshold value, described at least one look-up table comprises a plurality of pixel values that are divided into heat, warm and cold pixel value, and each pixel value all has a corresponding threshold value, and wherein said corresponding threshold value is looser with the value increase of described a plurality of pixel values.
7. module as claimed in claim 5, the described corresponding threshold value of each pixel value in the described corresponding threshold ratio redness of each pixel value of its Smalt and the green described a plurality of pixel values is looser.
8. module as claimed in claim 6, the described corresponding threshold value of the described green of described corresponding threshold ratio of some pixel value of wherein said redness is looser.
9. bad pixel detection module, it comprises:
Quick test bad pixel detection submodule, it can be operated so that difference and threshold value between good neighbour's pixel value of the current pixel value of the color in the image and described color are compared, and described threshold value is based on the human visual system response of described color is created the noticeable bad pixel of zero in described image; And
Complete test bad pixel detection submodule, whether it can operate being bad pixel in the described nuclear by the test fully of nuclear being assessed noticeable bad pixel.
10. module as claimed in claim 9, wherein said quick test bad pixel detection submodule can operate with described current pixel value and one and only previous good neighbour's pixel value after tested compare.
11. module as claimed in claim 10, wherein said quick test bad pixel detection submodule further can be operated to stop the test to the current pixel with described current pixel value under the situation of the scope that is positioned at described threshold value in described difference, and further can operate to begin the test to the new pixel with new pixel value, described new pixel value becomes described current pixel value.
Therefore 12. module as claimed in claim 11, wherein said quick test bad pixel detection submodule can be operated with the described scope that exceeds described threshold value in described difference and stop test to described current pixel when being noticeable bad pixel; And described complete test bad pixel detection submodule can be operated to contrast residual pixel neighbour in the described nuclear and restart test to described noticeable bad pixel.
13. module as claimed in claim 9, wherein said quick test bad pixel detection submodule further comprises at least one look-up table of described color, described color comprises a plurality of pixel values that are divided into heat, warm and cold pixel value, each pixel value all is mapped as a respective threshold, and wherein said respective threshold is looser with the value increase of described a plurality of pixel values.
14. module as claimed in claim 13, the described threshold ratio redness of each pixel value of its Smalt and green described threshold value are looser.
15. module as claimed in claim 14, the described threshold value of the described green of described threshold ratio of some pixel value of wherein said redness is looser.
16. a bad pixel detection module, it comprises:
Device for quick testing, it is used for by the difference between good neighbour's pixel value of the current pixel value of the color of image and described color and threshold value are compared current pixel being tested fast, and described threshold value is based on the human visual system response of described color is created the noticeable bad pixel of zero in described image; And
Complete testing apparatus, whether it is used for testing to assess described noticeable bad pixel fully to noticeable bad pixel is the bad pixel of nuclear.
17. module as claimed in claim 16, wherein said device for quick testing with described current pixel value and one and only previous good neighbour's pixel value after tested compare.
18. module as claimed in claim 17, wherein said device for quick testing comprises: stop device, it is used for stopping the test to described current pixel under described difference is positioned at the situation of scope of described threshold value; And starting apparatus, it is used to begin the test to the new pixel with new pixel value, and described new pixel value becomes described current pixel value.
Therefore 19. module as claimed in claim 18, wherein said termination device are included in that described difference exceeds the described scope of described max-thresholds and stop device to the test of described current pixel when being noticeable bad pixel; And wherein said complete testing apparatus comprises the device that is used to restart to the test of described noticeable bad pixel.
20. module as claimed in claim 16, it further comprises the device of the described threshold value that is used to search described color, wherein said color comprises a plurality of pixel shader that are divided into the heat that is stored at least one look-up table, warm and cold pixel value, and wherein each pixel value all is mapped as a respective threshold.
21. module as claimed in claim 20, the described threshold ratio redness of each pixel shader of its Smalt and green described threshold value are looser.
22. module as claimed in claim 21, the described threshold value of the described green of described threshold ratio of some pixel shader of wherein said redness is looser.
23. a method that is used for bad pixel detection, it may further comprise the steps:
By the difference between good neighbour's pixel value of the current pixel value of the color in the image and described color and threshold value are compared current pixel is tested fast, described threshold value is based on the human visual system response of described color is created the noticeable bad pixel of zero in described image; And
Whether test to assess described noticeable bad pixel fully to noticeable bad pixel is bad pixel in the nuclear.
24. method as claimed in claim 23, wherein said quick testing procedure comprise described current pixel value and one and step that previous good neighbour's pixel value after tested compares only.
25. method as claimed in claim 23, wherein said quick testing procedure may further comprise the steps: be positioned at the test that stops under the situation of scope of described threshold value the current pixel with described current pixel value in described difference; And beginning test to new pixel with new pixel value, described new pixel value becomes described current pixel value.
26. method as claimed in claim 25, the step of wherein said termination test comprise when described difference exceeds the described scope of described threshold value and therefore is noticeable bad pixel termination to the step of the test of described current pixel; And wherein said complete testing procedure comprises the step of restarting the test of described noticeable bad pixel.
27. method as claimed in claim 23, wherein said quick testing procedure comprises the step of searching described threshold value from least one look-up table, wherein said color comprises a plurality of pixel shader that are divided into the heat that is stored in described at least one look-up table, warm and cold pixel value, and wherein each pixel value is mapped as a respective threshold.
28. method as claimed in claim 27, the described threshold ratio redness of each pixel shader of its Smalt and green described threshold value are looser.
29. method as claimed in claim 26, the described threshold value of the described green of described threshold ratio of some pixel shader of wherein said redness is looser.
30. the program command that can carry out by processor, described program command when carrying out, can operate with:
By the difference between good neighbour's pixel value of the current pixel value of the color in the image and described color and threshold value are compared current pixel is tested fast, described threshold value is based on the human visual system response of described color is created the noticeable bad pixel of zero in described image; And
Whether test to assess described noticeable bad pixel fully to noticeable bad pixel is bad pixel in the nuclear.
31. program command as claimed in claim 30, wherein when carrying out, the described program command of described quick test further can operate with described current pixel value with one and only one before after tested good neighbour's pixel value compare.
32. program command as claimed in claim 31, wherein when carrying out, the described program command of described quick test further can be operated to stop the test to the current pixel with described current pixel value under the situation of the scope that is positioned at described threshold value in described difference, and further can operate to begin the test to the new pixel with new pixel value, described new pixel value becomes described current pixel value.
33. program command as claimed in claim 32, wherein when carrying out, therefore the described program command of described quick test further can be operated with the described scope that exceeds described threshold value in described difference and stop test to described current pixel when being noticeable bad pixel, and the described described program command of test fully further can be operated to restart the test to described noticeable bad pixel.
34. program command as claimed in claim 30, wherein when carrying out, the described program command of described quick test can be operated to search described threshold value from least one look-up table, wherein said color comprises a plurality of pixel values that are divided into the heat that is stored in described at least one look-up table, warm and cold pixel value, and wherein each pixel value is mapped as a respective threshold, wherein increase with described pixel value, described respective threshold becomes looser.
35. program command as claimed in claim 34, the described threshold ratio redness of each pixel value of its Smalt and green described threshold value are looser.
36. program command as claimed in claim 35, the described threshold value of the described green of described threshold ratio of some pixel value of wherein said redness is looser.
37. the program command that can carry out by processor, described program command when carrying out, can operate with:
By with the pixel value of the current pixel in the image and only the difference between the pixel value of good neighbour's pixel and threshold value compare described current pixel tested fast, described threshold value is to select to be used for to create in described image the noticeable bad pixel of zero.
38. program command as claimed in claim 37, wherein said program command further can be operated when carrying out, and whether be bad pixel in nuclear only described current pixel tested fully under not by the situation of described quick test at described current pixel if assessing described current pixel.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US76066906P | 2006-01-19 | 2006-01-19 | |
| US60/760,669 | 2006-01-19 | ||
| US11/506,350 | 2006-08-18 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101371566A true CN101371566A (en) | 2009-02-18 |
Family
ID=40414072
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2007800030768A Pending CN101371566A (en) | 2006-01-19 | 2007-01-19 | Method and apparatus for defective pixel detection based on the human visual system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101371566A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107077604A (en) * | 2014-10-20 | 2017-08-18 | 微软技术许可有限责任公司 | The skin of face mask generation detected for the rhythm of the heart |
-
2007
- 2007-01-19 CN CNA2007800030768A patent/CN101371566A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107077604A (en) * | 2014-10-20 | 2017-08-18 | 微软技术许可有限责任公司 | The skin of face mask generation detected for the rhythm of the heart |
| US10529071B2 (en) | 2014-10-20 | 2020-01-07 | Microsoft Technology Licensing, Llc | Facial skin mask generation for heart rate detection |
| CN107077604B (en) * | 2014-10-20 | 2020-07-21 | 微软技术许可有限责任公司 | Facial skin mask generation for cardiac rhythm detection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8971659B2 (en) | Bad pixel cluster detection | |
| CN103997631B (en) | A kind of LED pointwise position contour extraction method based on camera shooting and system | |
| US9805458B2 (en) | Method and system for detecting defective pixels and screen imperfections of a mobile device | |
| US8867817B1 (en) | Display analysis using scanned images | |
| US20140016005A1 (en) | Information processing apparatus, information processing method, and information processing program | |
| CN101277385A (en) | Method and apparatus to detect a dead pixel of an image sensor and method and apparatus to capture an image from an image sensor | |
| CN104966493A (en) | Method for compensating spliced bright and dark lines | |
| CN102017639B (en) | Methods and apparatuses for addressing chromatic aberrations and purple fringing | |
| CN106534821A (en) | Image processing device and auto white balancing method | |
| CN105306843A (en) | Dead pixel processing method and system for image sensor | |
| CN109068025A (en) | A kind of camera lens shadow correction method, system and electronic equipment | |
| CN104200775A (en) | LED defective pixel treatment method | |
| CN103369265A (en) | Pixel information management apparatus and image capture apparatus using the same | |
| WO2013175703A1 (en) | Display device inspection method and display device inspection device | |
| JP2006074734A (en) | Pixel testing method, method of correcting output voltage of pixel, device, program, and method for processing defective pixel, and recording medium with program recorded thereon | |
| CN101371566A (en) | Method and apparatus for defective pixel detection based on the human visual system | |
| US7920186B2 (en) | Method and apparatus for defective pixel detection based on the human visual system | |
| US20040212705A1 (en) | Method and apparatus for defect compensation in an image sensor | |
| CN104464634B (en) | LED calibration results inspection method and correlation-corrected method and device | |
| CN115346464A (en) | Display compensation data setting method, display compensation method and driving chip | |
| KR20150009842A (en) | System for testing camera module centering and method for testing camera module centering using the same | |
| CN114693656A (en) | Method for detecting display defects of LED display screen and method for calibrating camera and optical filter | |
| CN103558702B (en) | The detection method of color membrane substrates and pick-up unit | |
| US7330580B2 (en) | System and method for inspecting an LCD panel | |
| CN101472036A (en) | Image processing method capable of removing flaw and device applied the method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20090218 |
|
| C20 | Patent right or utility model deemed to be abandoned or is abandoned |