JP2004109788A - Method of generating false sem image data and method of examining defect of photomask - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of examining defects of a photomask, which has less restrictions and uses SEM image data superior in defect extraction precision and to provide a method of generating false SEM image data for use in the same. <P>SOLUTION: With respect to individual pixels of original image data based on a designed shape of a product, a ratio Ra of the number of pixels of a prescribed pattern part to the total number of pixels of a prescribed neighboring area including a pixel being an object and pixels around the pixel is obtained. The obtained ratio Ra is converted to a corresponding luminance value by a prescribed density display function F(Ra) having a ratio Ra preliminarily obtained correspondingly to the density display state of an SEM image, as a parameter. Image data is newly generated where the luminance value obtained by conversion is assigned to each corresponding pixel of original image data as its luminance value, and this image data is taken as false SEM image data. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method of generating pseudo SEM image data similar to SEM image data of a product, and a method of inspecting a photomask for defects using the pseudo SEM image.
[0002]
[Prior art]
2. Description of the Related Art In recent years, various LS1s represented by ASICs have been increasingly required to have higher integration and higher functionality due to the trend toward higher functionality and lighter and smaller electronic devices.
That is, it is required for an LSI such as an ASIC to realize a high function by reducing the chip size as much as possible.
An LSI such as the ASIC described above creates graphic data (also referred to as pattern data) for producing a photomask pattern through functions, logic design, circuit design, layout design, and the like. It is manufactured through a number of steps of transferring a pattern of a photomask onto a wafer by reduction projection exposure or the like and performing a semiconductor element manufacturing process.
In general, a photomask is disposed on a light-shielding film of a photomask substrate (also referred to as a photomask blank) using an electron beam exposure apparatus or a photo exposure apparatus such as an excimer wavelength using the above-described graphic data (pattern data). Exposure and drawing are performed on the provided photosensitive resist, and development and etching processes are performed.
That is, by applying an ionizing radiation to only a predetermined area by an exposure apparatus, on a photosensitive resist that has been coated and dried on a metal thin film of a photomask substrate provided with a light-shielding metal thin film on one surface of a glass substrate. After forming a latent image and developing the photosensitive resist to obtain a resist pattern of a desired shape corresponding to the irradiation area of ionizing radiation, the resist pattern is further used as an etching resistant resist, and the metal thin film is formed into a resist pattern. To obtain a photomask having a desired metal thin film pattern.
In the case where the pattern of the photomask is reduced and projected on the wafer and the pattern is transferred, the photomask is also called a reticle mask.
[0003]
As described above, the pattern of the photomask is transferred onto the wafer by reduction projection exposure or the like, and a circuit pattern is formed on the wafer. However, with the increasingly higher integration of LSl, the size of the exposure shape has recently been increased. Since the (exposure size on the wafer) has become finer and closer to or smaller than the wavelength of the exposure light, defect inspection using SEM image data (image data of an electron beam microscope) may be performed. It has become.
Defect inspection of a photomask using SEM image data is usually performed by comparing SEM image data.
When the same pattern is repeated on the photomask, defects are extracted by comparing SEM image data of the same pattern at different positions.
Alternatively, a defect is extracted by comparing SEM image data of a good product expected to be the same image data prepared in advance with target image data.
However, restrictions such as the condition that the same pattern is repeated on the photomask and preparation of a good SEM image are serious problems in practice, and are not compared with the design information. There was also a problem that the accuracy of extraction was low.
[0004]
Japanese Patent Application Laid-Open No. 5-258703 discloses an electron beam inspection method and system for comparing an X-ray mask or equivalent with an SEM image and data. No method of generating a pseudo-SEM image that can perform the comparison with good accuracy is disclosed.
[0005]
[Patent Document 1]
JP-A-5-258703 (column [0022] on page 6, column [0112] on page 15, FIG. 1)
[0006]
As described above, finer patterns and higher densities of photomasks have been further advanced, and defect inspection methods using SEM image data have been performed. A conventional method for inspecting a defect of a photomask using SEM image data, which compares data with each other, has limitations, and there is also a problem in terms of the accuracy of defect extraction.
The present invention is intended to provide a method for inspecting a defect of a photomask using SEM image data which corresponds to these, has few restrictions, and is excellent in accuracy of defect extraction. An object of the present invention is to provide a photomask inspection method for extracting a defect by comparing SEM image data of a target with pseudo SEM image data generated based on a design shape of a product.
At the same time, an object is to provide a method for generating pseudo SEM image data used in such a photomask inspection method.
[0007]
[Means for Solving the Problems]
The method of generating pseudo SEM image data of the present invention is a method of generating a pseudo SEM image similar to a SEM image of a product, and targets each pixel of the original image data based on the design shape of the product. The ratio Ra of the number of pixels of the predetermined picture portion to the total number of pixels in the predetermined neighborhood including the pixel and the pixels around the pixel is obtained, and the obtained Ra is obtained in advance corresponding to the light and dark display state of the SEM image. Is converted into a corresponding luminance value by a predetermined luminance display function F (Ra) using the ratio Ra as a parameter, and the luminance values obtained by the conversion are respectively converted into corresponding pixels of the original image data. Further, image data allocated as the luminance value is newly created and used as pseudo SEM image data.
In the above description, the product is a photomask, the predetermined pattern portion is a metal pattern portion, and represents the design shape of the product corresponding to the highlight position, which has been obtained in advance in accordance with the bright and dark display state of the SEM image. The ratio of the number of pixels of the predetermined picture portion to the total number of pixels in the vicinity area of the pixels of the image data is R1, and the brightness of the middle portion of the glass portion, the brightness of the highlight portion, and the brightness of the middle portion of the metal portion of the SEM image are , T1, T2, and T3, respectively, the predetermined luminance display function F (Ra) is such that Ra is 0 to 0. 5 in the range of brightness T1, R1 in the range of brightness T2, 1.. 0 and the luminance T3. Ra = 0.5 between 5 and R1. 5, represented by a straight line connecting the point of F (Ra) = T1 and the point of Ra = R1, F (Ra) = T2, and Ra is represented by R1-1. 0, Ra = R1, F (Ra) = T2 and Ra = 1. 0, F (Ra) = T3, represented by a straight line connecting the points.
Further, in the above, a predetermined neighboring area including the target pixel and pixels around the target pixel is a pixel area included within a certain radius around the target pixel. Is what you do.
Note that the luminance of the middle of the glass part and the luminance of the metal part of the SEM image referred to here are luminances that are not affected by the luminance characteristics of the boundary part between the glass part and the metal part of the SEM image, in other words. , Respectively, corresponds to the luminance at the center of a sufficiently large glass part of the photomask or at the center of a sufficiently large metal part.
[0008]
Alternatively, the method of generating pseudo SEM image data of the present invention is a method of generating a pseudo SEM image similar to an SEM image of a product, wherein each pixel of the original image data based on the design shape of the product is subjected to a target object. The luminance value Rb obtained by averaging the luminance values of all the pixels in a predetermined neighboring area including the pixel to be determined and the pixels around the pixel is obtained, and the obtained luminance value Rb is determined in advance in the light / dark display state of the SEM image. Is converted into a corresponding luminance value by a predetermined luminance display function G (Rb) using the luminance value Rb as a parameter, and the luminance values obtained by the conversion are respectively assigned to the original image. It is characterized in that image data newly allocated to each pixel corresponding to the data as its luminance value is newly created and used as pseudo SEM image data. In the above, the product is a photomask, the predetermined pattern portion is a metal pattern portion, and the position corresponding to the highlight position and the position of the metal pattern portion, which have been determined in advance corresponding to the light and dark display state of the SEM image, are determined. The corresponding luminance values Rb of the pixels of the original image data based on the design shape of the product are R11 and R21, respectively, and the luminance in the middle of the glass part, the luminance of the highlight part, and the luminance of the metal part in the SEM image. Are defined as T11, T21, and T31, respectively, the predetermined luminance display function G (Rb) is such that Rb is 0 to 0. In a range of 5 × R21, the luminance T1 is R1 and the luminance T21 is R21. Rb = 0.5 between 5 × R21 to R11. It is represented by a straight line connecting the point of 5 × R21, G (Rb) = T11 and the point of Rb = R11, G (Rb) = T21, and Rb = R11, G (Rb) = T21 when Rb is between R11 and R21. And a point connecting Rb = R21 and G (Rb) = T31.
The luminance value R11 of the pixel of the original image data based on the design shape of the product corresponding to the position of the metal picture portion here is a luminance which is not affected by the luminance characteristics of the boundary portion between the glass portion and the metal portion of the SEM image. Corresponds to the luminance at the center of a sufficiently large metal portion of the photomask.
Further, in the above, a predetermined neighboring area including the target pixel and pixels around the target pixel is a pixel area included within a certain radius around the target pixel. Is what you do.
[0009]
The photomask defect inspection method of the present invention is a photomask defect inspection method for extracting a defect from SEM image data of a product, and includes a SEM image data of a product and a pseudo-mask of the present invention corresponding to the SEM image data. The method is characterized in that a defective portion is extracted by comparing with pseudo SEM image data created by a method of generating SEM image data.
[0010]
[Action]
The pseudo-SEM image data generation method of the present invention, with such a configuration, enables generation of pseudo-SEM image data close to SEM image data of a product having a designed shape.
Specifically, for each pixel of the original image data based on the design shape of the product, the pixel of the predetermined pattern portion with respect to the total number of pixels of the predetermined neighborhood including the target pixel and the pixels around the pixel A ratio Ra of the number is obtained, and the obtained Ra is determined by a predetermined luminance display function F (Ra) having the ratio Ra as a parameter, which is obtained in advance corresponding to the light and dark display state of the SEM image. The image data is converted into a luminance value, and the luminance value obtained by the conversion is assigned to each corresponding pixel of the original image data as the luminance value, and image data is newly created to be pseudo SEM image data. Or, for each pixel of the original image data based on the design shape of the product, all the pixels in the predetermined neighboring area including the target pixel and the neighboring pixels adjacent to the pixel are brightened. A luminance value Rb is obtained by averaging the values, and the obtained luminance value Rb is obtained in advance corresponding to the light and dark display state of the SEM image. A predetermined luminance display using the luminance value Rb as a parameter A function G (Rb) is used to convert to a corresponding brightness value, and the brightness value obtained by the conversion is assigned to each corresponding pixel of the original image data as the brightness value, thereby newly creating image data. This is achieved by using pseudo SEM image data.
The predetermined neighboring area including the target pixel and pixels around the target pixel includes a pixel area included within a certain radius around the target pixel.
[0011]
By adopting such a configuration, the defect inspection method for a photomask of the present invention can provide a defect inspection method for a photomask using SEM image data which has few restrictions and is excellent in the accuracy of defect extraction. I have.
Specifically, the SEM image data of the product to be inspected can be compared with the pseudo SEM image data generated based on the original image data based on the design shape of the product, that is, the inspection object is directly compared with the design information. By doing so, it is possible to further improve the accuracy of defect extraction.
When the design data (design pattern data) is represented by a figure code expression such as a polygon, rectangle, or trapezoid, original image data can be obtained by converting the whole or a part of the design data into a raster format.
[0012]
BEST MODE FOR CARRYING OUT THE INVENTION
An embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a processing flowchart of an embodiment of a method for generating pseudo SEM image data according to the present invention. FIG. 2A is a diagram showing a monitor display state of design data, and FIG. FIG. 2C is a diagram showing a monitor display state of the SEM image. FIG. 2C shows each position and its position when the pixel at the boundary is viewed from the A1 side to the A2 side at the A1-A2 position in FIG. FIG. 3 is a diagram showing a function F (Ra), FIG. 4 is a diagram showing a function G (Rb), and FIG. 5 is a photomask of the present invention. FIG. 6 is a processing flowchart of an example of an embodiment of the defect inspection method of FIG.
Note that S11 to S24 and S171 to S191 in FIG. 1 and S41 to S49 in FIG. 5 are processing steps.
2 and 3, reference numeral 110 denotes a monitor display state of design data, 111 denotes a picture part, 112 denotes a non-picture part, 120 denotes a monitor display state of a SEM image of a product, 121 denotes a picture part (also referred to as a metal layer part), 122 is a non-picture part (also called a glass part), 125 is a highlight part, and T1, T2, T3, T11, T21 and T31 are luminance values.
[0013]
First, a first example of an embodiment of a method for generating pseudo SEM image data according to the present invention will be described.
This example is a method of generating pseudo SEM image data similar to the SEM image data of a photomask product. In brief, each pixel of the original image data based on the design shape of the photomask product has a target pixel and a target pixel. The ratio Ra of the number of pixels of the predetermined pattern portion to the total number of pixels in the predetermined vicinity area including the pixels around the pixel is obtained, and the obtained ratio Ra is obtained in advance corresponding to the light and dark display state of the SEM image. The luminance value is converted into a corresponding luminance value by a predetermined luminance display function F (Ra) using the ratio Ra as a parameter, and the luminance values obtained by the conversion are respectively assigned to the corresponding pixels of the original image. Image data newly assigned as the luminance value is newly created and used as pseudo SEM image data.
First, a method of setting the luminance display function F (Ra) will be briefly described.
In the monitor display state 110 of the photomask design data, as shown in FIG. 2A, the boundary between the picture part 111 and the non-picture part 112 is clearly displayed, but the corresponding SEM image of the photomask product is displayed. It is known that, as shown in FIG. 2B, a highlight portion 125 is generated along the boundary between the picture portion 121 and the non-picture portion 122.
In the SEM image shown in FIG. 2B, when the pixels at the boundary between the picture portion and the non-picture portion are viewed in order from the A1 side to the A2 side at the A1-A2 position, each position and the pixel at that position are considered. Is as shown in FIG. 2C.
For this reason, it is necessary to consider the characteristics shown in FIG. 2C when producing pseudo SEM image data used for comparison with SEM image data of a photomask product.
In the case of this example, the ratio Ra of the number of pixels in the predetermined picture portion to the total number of pixels in the predetermined neighborhood including the target pixel and the pixels surrounding the target pixel is set so that each pixel can satisfy such characteristics. Using the function F (Ra) shown in FIG. 3 with the parameter as a parameter, this corresponds to the characteristic shown in FIG.
As a result, the brightness change at the boundary between the picture portion and the non-picture portion of the image of the pseudo SEM image data to be produced can be made close to the SEM image of the product.
[0014]
Hereinafter, the processing flow will be briefly described with reference to FIG.
The design data (S11) is displayed on a monitor. (S12)
At the time of monitor display, the design data is divided into pixels. The monitor here is displayed in a predetermined pixel unit. The total number of pixels is N, and there are first to Nth pixels. The Pith pixel is assumed to be Pi.
The range of the neighborhood area is determined in advance (S13), first, i = 1 (S14), and all the pixels in the neighborhood area are extracted for the pixel Pi (S16). The ratio Ra of the number of pixels is calculated. (S17)
The predetermined neighboring region including the target pixel and pixels around the target pixel includes, but is not limited to, a pixel region included within a certain radius around the target pixel.
Alternatively, for example, a range of (2n + 1) pixels × (2n + 1) pixels (n is an integer) around the target pixel can be cited as the neighboring area.
Next, the luminance value of the pixel Pi is obtained from the calculated Ra using the function F (Ra) shown in FIG. (S18-S19)
By a conversion using the function F (Ra), a luminance value corresponding to the calculated Ra is obtained.
Then, the obtained luminance value of the pixel Pi is stored in the memory. (S20)
Next, the processes of S15 to S20 are sequentially performed on the second and subsequent pixels, and the brightness of the monitor display is obtained for each pixel, and the obtained brightness value is stored in the memory in association with the pixel. (S21-S22-S15-S20)
After performing for all pixels, monitor display of all pixels is performed using the obtained luminance value corresponding to each pixel stored in the memory as necessary. (S23)
In the above processing, data processing can be performed at any stage without displaying an image at all, but it is effective to display on a screen or print out on paper at each stage. is there.
[0015]
Next, a second example of the embodiment of the method for generating pseudo SEM image data of the present invention will be described.
The second example is also a method of generating pseudo SEM image data similar to the SEM image data of a photomask product. In brief, each pixel of the original image data based on the design shape of the product has a target pixel and a target pixel. A luminance value Rb obtained by averaging the luminance values of all the pixels in a predetermined neighboring area including the pixel around the pixel is obtained, and the obtained luminance value Rb is previously determined in accordance with the bright and dark display state of the SEM image. The obtained luminance value Rb is converted into a corresponding luminance value by a predetermined luminance display function G (Rb) using the luminance value Rb as a parameter, and the luminance values obtained by the conversion are respectively converted into the corresponding luminance values of the original image. Image data newly allocated to the pixel as its luminance value is newly created and used as pseudo SEM image data.
The method of setting the luminance display function G (Rb) is also performed in consideration of the characteristics shown in FIG. 2C as in the first example. However, unlike the first example, the display monitor G (Rb) is different from the first example. The luminance value corresponding to each pixel is used, and for the pixel of interest, a value obtained by averaging the luminance values of all the pixels in the vicinity area is newly set as the luminance value for display.
[0016]
The processing flow is such that in the flow of the first example, processing steps S171 to S191 to S20 are performed instead of the processing steps S17 to S20.
After the pixels in the neighboring area are determined for the target pixel Pi (S15) (S16), the luminance values of all the pixels in the neighboring area are averaged, and the average luminance Rb is calculated. (S171)
Next, using the function G (Rb) shown in FIG. 4, the luminance value of the pixel Pi is obtained from the calculated Rb. (S181 to S191)
By a conversion using the function F (Rb), a luminance value corresponding to the calculated Rb is obtained.
Then, the obtained luminance value of the pixel Pi is stored in the memory. (S20)
Then, as in the first example, after performing for all the pixels, monitor display of all the pixels is performed using the obtained luminance value corresponding to each pixel stored in the memory as necessary. (S23)
In this example, in each process, data processing can be performed at any stage without displaying an image at all. However, each process can be displayed on a screen or printed out on paper. It is effective to do.
[0017]
Next, an example of an embodiment of a photomask defect inspection method according to the present invention will be described. This example uses an SEM image of a photomask product and a pseudo SEM image generation method according to the first or second example. This is a photomask defect inspection method for extracting a defective portion by comparing the produced pseudo SEM image.
Hereinafter, this example will be described with reference to FIG.
First, the SEM image of the product and the corresponding image of the pseudo SEM image data are displayed on the same monitor or a monitor of the same performance, respectively.
At this stage, the luminance value corresponding to each pixel is stored in a predetermined memory unit.
Next, i = 1, the luminance value T1i of the i-th pixel of the SEM image of the product and the luminance value T2i of the corresponding pixel of the pseudo image are extracted, and the luminance values T1i and T2i are compared (S42, S43). , S44), the quality of the i-th product pixel is determined. (S45)
Then, the determination result is stored in the memory. (S46)
Next, the processing of S42 to S46 is sequentially performed on each pixel of the SEM images of the second and subsequent products, and the quality of each pixel is determined for each pixel, and the result is stored in a memory in association with the pixel. . (S42, S43, S44-S46)
After performing the process for all the pixels of the SEM image of the product, the defective portion is extracted in consideration of the continuity and the density of the defective pixels by using the determination result corresponding to each pixel stored in the memory. (S49)
Whether the pixel is good or bad is determined, for example, based on whether the difference between the luminance value T1i and the luminance value T2i is equal to or less than a predetermined value Ka, or whether T1i / T2i is equal to or more than a predetermined value kb.
[0018]
【The invention's effect】
As described above, the present invention makes it possible to provide a method for generating pseudo SEM image data that can generate an SEM image close to the SEM image of a product having a designed shape, and at the same time, has less restrictions and is more accurate in terms of defect extraction. It is possible to provide a photomask defect inspection method using excellent SEM image data.
[Brief description of the drawings]
FIG. 1 is a process flowchart of an embodiment of a method for generating pseudo SEM image data according to the present invention.
2A is a view showing a monitor display state of design data, FIG. 2B is a view showing a monitor display state of an SEM image of a product, and FIG. 2C is a view showing FIG. FIG. 6B is a diagram illustrating a relationship between each position and the luminance value of the pixel at the position when the pixels at the boundary at the A1-A2 position in FIG.
FIG. 3 is a diagram showing a function F (Ra).
FIG. 4 is a diagram showing a function G (Rb).
FIG. 5 is a processing flowchart of an example of an embodiment of a photomask defect inspection method according to the present invention.
[Explanation of symbols]
110 Monitor display state of design data 111 Picture part 112 Non-pattern part 120 Monitor display state of SEM image of product 121 Picture part (also called metal layer part)
122 Non-picture part (also called glass part)
125 Highlight parts T1, T2, T3 Luminance values T11, T21, T31 Luminance values

Claims (7)

A method for generating a pseudo SEM image similar to a SEM image of a product, wherein each pixel of the original image data based on the design shape of the product includes a predetermined pixel including a target pixel and pixels around the pixel. A ratio Ra of the number of pixels of the predetermined pattern portion to the total number of pixels in the vicinity area is determined, and the determined Ra is determined in advance corresponding to the light and dark display state of the SEM image. Is converted into a corresponding luminance value by the luminance display function F (Ra), and the luminance data obtained by the conversion is assigned to each corresponding pixel of the original image data as the luminance value. A method for generating pseudo SEM image data, which is newly created and used as pseudo SEM image data. 2. The product according to claim 1, wherein the product is a photomask, the predetermined picture portion is a metal picture portion, and the design shape of the product corresponding to the highlight position, which has been obtained in advance corresponding to the light and dark display state of the SEM image. The ratio of the number of pixels of the predetermined picture portion to the total number of pixels in the vicinity area of the pixels of the image data is R1, and the brightness of the middle of the glass portion, the brightness of the highlight portion, and the brightness of the middle of the metal portion of the SEM image are , T1, T2, and T3, respectively, the predetermined luminance display function F (Ra) is such that Ra is 0 to 0. In the range of # 5, the brightness T1, R1 and the brightness T2, 1.. The brightness T3 is set at 0, and Ra is set to 0. Ra = 0.5 between R5 and R1. 5, represented by a straight line connecting a point of F (Ra) = T1 and a point of Ra = R1, F (Ra) = T2, where Ra is R1 to 1. Between 0, Ra = R1, F (Ra) = T2 and Ra = 1. A method for generating pseudo SEM image data, characterized by being represented by a straight line connecting points of 0, F (Ra) = T3. 3. A method according to claim 1, wherein a predetermined neighboring area including the target pixel and pixels around the target pixel is a pixel area included within a certain radius around the target pixel. A method for generating pseudo SEM image data. A method for generating a pseudo SEM image similar to a SEM image of a product, wherein each pixel of the original image data based on the design shape of the product includes a predetermined pixel including a target pixel and pixels around the pixel. The brightness value Rb obtained by averaging the brightness values of all the pixels in the vicinity area is obtained, and the obtained brightness value Rb is obtained in advance in accordance with the light and dark display state of the SEM image. According to a predetermined luminance display function G (Rb) as a parameter, the luminance value is converted into a corresponding luminance value, and the luminance value obtained by the conversion is assigned to each corresponding pixel of the original image data as the luminance value. , A method of generating pseudo SEM image data, wherein image data is newly created and used as pseudo SEM image data. 5. The product according to claim 4, wherein the product is a photomask, the predetermined pattern portion is a metal pattern portion, and the position corresponding to the highlight position and the position of the metal pattern portion, which have been determined in advance corresponding to the light and dark display state of the SEM image. The corresponding luminance values Rb of the pixels of the original image data based on the design shape of the product are R11 and R21, respectively, and the luminance in the middle of the glass part, the luminance of the highlight part, and the luminance of the metal part in the SEM image. Are defined as T11, T21, and T31, respectively, the predetermined luminance display function G (Rb) is such that Rb is 0 to 0. In a range of 5 × R21, luminance T1, luminance R21 is represented by luminance T21, and luminance R21 is represented by luminance T31. Rb = 0.5 between 5 × R21 to R11. It is represented by a straight line connecting the point of 5 × R21, G (Rb) = T11 and the point of Rb = R11, G (Rb) = T21, and Rb = R11, G (Rb) = T21 when Rb is between R11 and R21. A pseudo SEM image data generation method characterized by being represented by a straight line connecting a point of Rb = R21 and a point of G (Rb) = T31. 6. A method according to claim 4, wherein the predetermined neighboring area including the target pixel and pixels surrounding the target pixel is a pixel area included within a certain radius around the target pixel. A method for generating pseudo SEM image data. 7. A defect inspection method for a photomask for extracting defects from an SEM image of a product, wherein the SEM image data of the product and pseudo SEM image data produced by the method according to claim 1 corresponding to the SEM image data. A defect inspection method for a photomask, wherein a defective portion is extracted by comparing

Images