JP7459007B2 - Defect inspection equipment and defect inspection method - Google Patents

Description

The present invention relates to a defect inspection apparatus and a defect inspection method for inspecting a pattern formed on a sample for defects.

In the manufacturing process of semiconductor devices, a circuit pattern is transferred onto a semiconductor substrate by reduction exposure using an exposure device (also called a "stepper" or "scanner"). In order to transfer a circuit pattern onto a semiconductor substrate (hereinafter also referred to as a ``wafer''), an exposure device uses a mask (also called a ``reticle'') on which an original image pattern (hereinafter also simply referred to as a ``pattern'') is formed. ) is used.

For example, cutting-edge devices require the formation of circuit patterns with line widths of several nanometers. As circuit patterns become finer, the original patterns on masks are also becoming finer. For this reason, a mask defect inspection apparatus is required to have high defect detection performance that can handle fine original patterns.

Defect inspection methods include the D-DB (Die to Database) method, which compares an inspection image based on an image captured by a defect inspection device with a reference image based on design data, and the D-DB (Die to Database) method, which compares an inspection image based on an image captured by a defect inspection device with a reference image based on design data. There is a DD (Die to Die) method that compares multiple regions.

The dimensions of a pattern formed on a mask may shift due to the pattern formation (writing) process. In the case of the D-DB method, when a dimension shift occurs, the edge position of the pattern in the inspection image no longer matches the reference image. If a positional deviation occurs between the reference image and the inspection image, the positional deviation location may be detected as a pseudo defect.

To accommodate the dimensional shift, when creating a reference image, corrections such as resizing to move the edge positions of the pattern and corner rounding to round the corners of the pattern are applied to the binary or multi-value developed image generated based on the design data.

For example, Patent Document 1 discloses a technique for determining a resizing amount in a resizing process from the distance between adjacent figures existing in the vicinity of each pixel in a developed image.

JP2006-208340

The dimensional shift of the pattern in the inspection image also depends on the size and shape of the pattern. Therefore, if uniform resizing processing is performed on various patterns, it is difficult to match the reference image and the inspection image for all patterns.

The present invention has been made in view of these points. That is, an object of the present invention is to provide a defect inspection device and a defect inspection method that can generate reference image data with different resizing amounts based on the size and shape of a pattern.

According to a first aspect of the present invention, a defect inspection apparatus includes a sample imaging mechanism, an image acquisition circuit that generates an inspection image based on image data of the sample imaged by the imaging mechanism, and a developed image from design data. The pattern is classified into classes from the generated expansion circuit and the expanded image, and the resizing amount is given to the edge of the pattern based on the resizing amount set for each classified class, and the resizing amount is shown as the resizing amount of each pixel. Create a map, execute adjacent pixel maximization processing of the resizing amount map, execute smoothing processing of the resizing amount map after the adjacent pixel maximization processing, execute dilation processing of the resizing amount map after the smoothing processing, and perform dilation A reference image that includes a resizing processing circuit that executes resizing processing that moves the position of the edge of the pattern pixel by pixel based on a resizing amount map after processing, and that generates a reference image using the expanded image after the resizing processing. It includes a generation circuit and a comparison circuit that compares the inspection image and the reference image. In the adjacent pixel maximization process, the resizing processing circuit calculates, along the edge of the pattern, a resizing amount of a second pixel adjacent to any of the first pixels for which a first predetermined value or more is set as the resizing amount. If the amount of resizing of the first pixel is larger than the amount of resizing of the first pixel, the amount of resizing of the first pixel is replaced with the amount of resizing of the second pixel. In the dilation process, the resizing processing circuit performs a resizing process such that the resizing amount is a third predetermined value among adjacent pixels other than the edge of the pattern on the first pixel for which the resizing amount is set to a second predetermined value or more. A resizing amount based on the resizing amount of the first pixel is given to the third pixel.

According to the first aspect of the present invention , in classifying patterns, the resizing processing circuit calculates a first distance from a pixel of interest provided on at least a first edge of the pattern to another pattern facing the pattern; a second distance from to a second edge opposite to the pixel of interest in the pattern, a third distance from the first edge where the pixel of interest is placed to the first end of the first edge, and a third distance from the first edge where the pixel of interest is placed, Classification of the pattern is performed based on the fourth distance from the first edge to the second end of the first edge, and in the smoothing process, pixels within a preset range centered on the pixel of interest are It is preferable to filter the resizing amount based on a Gaussian function based on the distance from the pixel and a weight of a power of the resizing amount of the pixel .

According to the first aspect of the present invention, it is preferable that the resizing processing circuit performs a process of subtracting the resizing amount applied to the resizing process from the resizing amount map after the resizing process on a pixel by pixel basis .

According to the first aspect of the present invention, the resizing processing circuit repeats the dilation processing, the resizing processing, and the subtraction processing until the resizing amount in the resizing amount map after the subtraction processing becomes equal to or less than a preset threshold. preferable.

According to the second aspect of the present invention, a defect inspection method includes the steps of: imaging a sample to generate an inspection image; generating a developed image from design data; and classifying patterns from the developed image. Then, the process of setting the resizing amount for each classified class, assigning the resizing amount to the edge of the pattern, creating a resizing amount map showing the resizing amount of each pixel , and maximizing the adjacent pixels of the resizing amount map. a step of executing a smoothing process on the resize amount map after the adjacent pixel maximization process, a step of executing an expansion process on the resize amount map after the smoothing process, and a process based on the resize amount map after the expansion process. , a step of executing resizing processing to move the edge position of the pattern pixel by pixel , a step of generating a reference image based on the expanded image after the resizing processing, and a step of comparing the inspection image and the reference image to perform inspection. and a step of performing the process. In the step of executing adjacent pixel maximization processing, the resizing amount of a second pixel adjacent to any of the first pixels for which a resizing amount that is equal to or greater than a first predetermined value is set along the edge of the pattern is determined as follows. If the resizing amount is larger than the resizing amount of one pixel, the resizing amount of the first pixel is replaced with the resizing amount of the second pixel. In the step of executing the expansion process, among the adjacent pixels other than the edge of the pattern on the first pixel for which the resizing amount is set to a second predetermined value or more, the first pixel whose resizing amount is a third predetermined value is selected. A resizing amount based on the resizing amount of the first pixel is given to the three pixels.

According to the second aspect of the present invention, in the step of classifying the pattern, the first distance from the pixel of interest provided on at least the first edge of the pattern to another pattern facing the pattern, and the distance from the pixel of interest to the pattern a second distance from the first edge on which the pixel of interest is placed to the first edge of the first edge; and a first edge on which the pixel of interest is placed. Classification of the pattern is performed based on the fourth distance from to the second end of the first edge, and in the step of performing smoothing processing, , the resizing amount is preferably filtered based on a Gaussian function based on the distance from the pixel of interest and a weight of a power of the resizing amount of the pixel .

According to the second aspect of the present invention, the defect inspection method preferably further includes the step of subtracting the resizing amount applied to the resizing process from the resizing amount map after the resizing process , pixel by pixel .

According to the second aspect of the present invention, the defect inspection method preferably repeats the expansion process, the resize process, and the subtraction until the resize amount in the resize amount map after the subtraction becomes equal to or less than a preset threshold.

According to the defect inspection device and the defect inspection method of the present invention, the defect inspection device can generate reference image data with different resizing amounts based on the size and shape of the pattern.

FIG. 1 is a diagram showing the overall configuration of a defect inspection apparatus according to an embodiment. FIG. 2 is a flowchart of an inspection process in a defect inspection apparatus according to an embodiment. FIG. 3 is a diagram illustrating an example of an actual image contour position in the defect inspection apparatus according to an embodiment. FIG. 4 is a flowchart of resizing processing in the defect inspection apparatus according to one embodiment. FIG. 5 is a diagram showing a pixel of interest on an edge of a line pattern corresponding to class A in a defect inspection apparatus according to an embodiment. FIG. 6 is a diagram showing a pixel of interest on an edge of a line pattern corresponding to class B in the defect inspection apparatus according to an embodiment. FIG. 7 is a diagram showing a pixel of interest on an edge of a line pattern corresponding to class C in the defect inspection apparatus according to an embodiment. FIG. 8 is a diagram showing a pixel of interest on an edge of a line pattern corresponding to class D in the defect inspection apparatus according to an embodiment. FIG. 9 is an example of a resizing amount map in the defect inspection apparatus according to one embodiment. FIG. 10 is a diagram illustrating an example of adjacent pixel MAX processing in the defect inspection apparatus according to an embodiment. FIG. 11 shows an example of a resize amount map after adjacent pixel MAX processing in a defect inspection apparatus according to an embodiment. FIG. 12 is an example of a resizing amount map after filter processing in the defect inspection apparatus according to one embodiment. FIG. 13 is a conceptual diagram for explaining expansion processing in the X direction in the defect inspection apparatus according to one embodiment. FIG. 14 is a diagram illustrating an example of expansion processing in the X direction in the defect inspection apparatus according to an embodiment. FIG. 15 is a conceptual diagram for explaining expansion processing in the Y direction in the defect inspection apparatus according to an embodiment. FIG. 16 is a diagram illustrating an example of expansion processing in the Y direction in the defect inspection apparatus according to an embodiment. FIG. 17 is a diagram illustrating an example of resizing processing in the defect inspection apparatus according to an embodiment. FIG. 18 is a diagram illustrating an example of repetition of expansion processing, resizing processing, and subtraction processing of a resize amount map in the defect inspection apparatus according to an embodiment. FIG. 19 is a diagram illustrating an example of extraction of contour points of a developed image in the defect inspection apparatus according to an embodiment.

The following describes the embodiments with reference to the drawings. The embodiments exemplify devices and methods for embodying the technical ideas of the invention. The drawings are schematic or conceptual, and the dimensions and ratios of each drawing are not necessarily the same as those in reality. The technical ideas of the present invention are not specified by the shapes, structures, arrangements, etc. of the components.

In the following, an electron beam image (hereinafter also referred to as "SEM image") of a pattern to be measured will be described using a scanning electron microscope (hereinafter referred to as "SEM (Scanning Electron Microscope)") as a sample defect inspection device. A defect inspection device that captures images will be described. Note that the defect inspection apparatus may use an optical microscope to capture an optical image of the pattern, or may use a light receiving element to capture an optical image of light reflected or transmitted through the sample. In addition, in this embodiment, a case will be explained in which the sample to be inspected is a mask, but the sample may be a sample with a pattern on its surface, such as a wafer or a substrate used for a liquid crystal display device. Bye.

1. Overall Configuration of Defect Inspection Apparatus First, an example of the overall configuration of the defect inspection apparatus will be described using FIG. 1. FIG. 1 is a diagram showing the overall configuration of a defect inspection apparatus 1. As shown in FIG.

As shown in FIG. 1, the defect inspection apparatus 1 includes an imaging mechanism 10 and a control mechanism 20.

The imaging mechanism 10 includes a sample chamber 11 and a lens barrel 12. The lens barrel 12 is installed above the sample chamber 11. For example, the lens barrel 12 has a cylindrical shape extending perpendicularly to the sample chamber 11. The surfaces of the sample chamber 11 and the lens barrel 12 that contact each other are open. The space formed by the sample chamber 11 and the lens barrel 12 is maintained in a vacuum (reduced pressure) state using a turbo molecular pump or the like.

Inside the sample chamber 11, a stage 13, a stage drive mechanism 14, and a detector 15 are provided.

A sample (mask) 300 is placed on the stage 13. The stage 13 is movable in the X direction parallel to the surface of the stage 13, and in the Y direction parallel to the surface of the stage 13 and intersecting the X direction. The stage 13 may also be movable in the Z direction perpendicular to the surface of the stage 13, or may be rotatable around a rotation axis on the XY plane with the Z direction as the rotation axis.

The stage drive mechanism 14 has a drive mechanism for moving the stage 13 in the X direction and the Y direction. Note that the stage drive mechanism 14 may have, for example, a mechanism that moves the stage 13 in the Z direction, or a mechanism that rotates the stage 13 around the rotation axis on the XY plane with the Z direction as the rotation axis. may have.

The detector 15 detects secondary electrons, reflected electrons, etc. emitted from the sample. The detector 15 transmits signals such as detected secondary electrons or reflected electrons, ie, SEM image data, to the image acquisition circuit 213.

Inside the lens barrel 12, an electron gun 16 and an electron optical system 17, which are components of the SEM, are provided.

The electron gun 16 is installed to emit an electron beam toward the sample chamber 11.

The electron optical system 17 converges and irradiates the sample 300 with the electron beam emitted from the electron gun 16 at a predetermined position. For example, the electron optical system 17 includes a plurality of focusing lenses 101 and 102, a plurality of scanning coils 103 and 104, and an objective lens 105. The electron beam emitted from the electron gun 16 is accelerated and then focused as an electron spot on the surface of the sample 300 placed on the stage 13 by the focusing lenses 101 and 102 and the objective lens 105. Scanning coils 103 and 104 control the position of the electron spot on sample 300.

The control mechanism 20 includes a control circuit 21, a storage device 22, a display device 23, an input device 24, and a communication device 25.

The control circuit 21 controls the entire defect inspection device 1. More specifically, the control circuit 21 controls the imaging mechanism 10 to acquire an SEM image. The control circuit 21 also controls the control mechanism 20 to compare the generated reference image with the inspection image and detect defects. That is, the control circuit 21 is a processor for performing defect inspection. For example, the control circuit 21 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), and a ROM (Read Only Memory), which are not shown. For example, the CPU expands a program stored in the ROM or the storage device 22 as a non-temporary storage medium into the RAM. Then, the control circuit 21 interprets and executes the program expanded in the RAM by the CPU to control the defect inspection device 1. The control circuit 21 may be, for example, a CPU device such as a microprocessor, or a computer device such as a personal computer. The control circuit 21 may also include a dedicated circuit (dedicated processor) in which at least some of the functions are performed by other integrated circuits, such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a graphics processing unit (GPU).

The control circuit 21 includes an expansion circuit 211, a reference image generation circuit 212, an image acquisition circuit 213, and a comparison circuit 214. Note that these may be configured by programs executed by integrated circuits such as a CPU, ASIC, FPGA, or GPU, or may be configured by hardware or firmware included in these integrated circuits, or by the hardware or firmware included in these integrated circuits. It may also be constituted by separate circuits controlled by an integrated circuit. In the following, a case will be described in which the control circuit 21 realizes the functions of the expansion circuit 211, the reference image generation circuit 212, the image acquisition circuit 213, and the comparison circuit 214 by a program executed.

The expansion circuit 211, for example, expands the design data held in the storage device 22 into data for each figure, and interprets the figure code indicating the figure shape of the figure data, figure dimensions, etc. Then, the expansion circuit 211 converts the design data into a binary or multivalued (e.g. 8-bit) image (hereinafter referred to as a "CAD image") as a pattern arranged in a grid with a predetermined quantization dimension as a unit. (also referred to as "expanded image"). The expansion circuit 211 calculates the occupancy rate occupied by the figure for each pixel of the expanded image. The figure occupancy rate within each pixel calculated in this way is the pixel value. For example, when a developed image is represented by 8-bit gradation data, the pixel value of each pixel is represented by a gradation value of 0 to 255. When the pixel value is 0, the figure occupancy rate is 0%, and when the pixel value is 255, the figure occupancy rate is 100%.

The reference image generation circuit 212 performs resizing processing and corner rounding processing on the developed image. Then, the reference image generation circuit 212 extracts a contour from the developed image after the resizing process and the corner rounding process, and generates a reference image (contour image). The reference image generation circuit 212 transmits the generated reference image to the comparison circuit 214 and the storage device 22. The reference image generation circuit 212 includes a resizing processing circuit 215 and a corner rounding processing circuit 216.

The resizing processing circuit 215 executes resizing processing. More specifically, the resizing processing circuit 215 of this embodiment classifies patterns, and gives pixels including each pattern (edge) a resize amount according to the class classification. The resizing processing circuit 215 generates a resizing amount map representing the resizing amount for each pixel. Then, the resizing processing circuit 215 moves the edge position of the pattern in the developed image based on the resizing amount map.

The corner rounding processing circuit 216 executes corner rounding processing for rounding the corner portion of each pattern after resizing processing.

The image acquisition circuit 213 acquires SEM image data from the detector 15 of the imaging mechanism 10. The image acquisition circuit 213 extracts a contour from the SEM image and generates an inspection image (contour image).

The comparison circuit 214 aligns the test image and the reference image, and calculates the amount of shift of the test image with respect to the reference image. Further, the comparison circuit 214 measures the amount of distortion of the inspection image based on, for example, variations in the amount of shift within the surface of the sample 300, and calculates a distortion coefficient. The comparison circuit 214 compares the test image and the reference image using an appropriate algorithm that takes into account the shift amount and distortion coefficient. If the error between the inspection image and the reference image exceeds a preset value, the comparison circuit 214 determines that there is a defect at the corresponding coordinate position of the sample 300.

The storage device 22 stores data and programs related to defect inspection. For example, the storage device 22 stores design data 221, parameter information 222 of inspection conditions, and inspection data 223. More specifically, for example, the parameter information 222 of inspection conditions includes the imaging conditions of the imaging mechanism 10, reference image generation conditions (classification information of resizing processing, etc.), and defect detection conditions. The inspection data 223 includes image data (unfolded image, reference image, SEM image, and inspection image), as well as data related to detected defects (coordinates, size, etc.). The storage device 22 also stores a defect inspection program 224 as a non-transient storage medium. The defect inspection program 224 is a program for causing the control circuit 21 to execute defect inspection.

Note that the storage device 22 may include various storage devices such as a magnetic disk storage device (HDD: Hard Disk Drive) or a solid state drive (SSD) as external storage. Further, the storage device 22 may include, for example, a drive for reading a program stored on a CD (Compact Disc) or a DVD (Digital Versatile Disc) as a non-temporary storage medium.

The display device 23 includes, for example, a display screen (for example, an LCD (Liquid Crystal Display) or an EL (Electroluminescence) display). The display device 23 displays, for example, defect detection results under the control of the control circuit 21.

The input device 24 is an input device such as a keyboard, a mouse, a touch panel, or a button switch.

The communication device 25 is a device for connecting to a network in order to send and receive data to and from an external device. Various communication standards may be used for communication. For example, the communication device 25 receives design data from an external device, and transmits defect inspection results and the like to the external device.

2. Overall Flow of Inspection Process Next, an example of the overall flow of the inspection process will be described using FIG. 2. FIG. 2 is a flowchart of the inspection process.

As shown in FIG. 2, the inspection process roughly includes an inspection image acquisition process (step S1), a reference image generation process (step S2), and a comparison process (step S3).

2.1 Inspection Image Acquisition Step First, an example of the inspection image acquisition step in step S1 will be described. The image acquisition circuit 213 acquires an SEM image of the sample 300 from the imaging mechanism 10 (step S11).

Next, the image acquisition circuit 213 performs filter processing to remove noise from the acquired SEM image (step S12).

Next, the image acquisition circuit 213 extracts the contour of the pattern from the filtered SEM image (step S13) and generates an inspection image (contour image). More specifically, the image acquisition circuit 213 extracts, for each SEM image, a plurality of contour positions (actual image contour positions) of each figure pattern in the SEM image as a real image (inspection image).

An example of extracting the actual image contour position in the image acquisition circuit 213 will be described below. FIG. 3 is a diagram showing an example of an actual image contour position.

The image acquisition circuit 213 performs differential filter processing to differentiate each pixel in the X direction and the Y direction using a differential filter such as a Sobel filter, and synthesizes primary differential values in the X direction and the Y direction. Then, the peak position of the profile using the synthesized first-order differential value is extracted as the contour position on the contour line (actual image contour line). The example in FIG. 3 shows a case where the contour position is extracted one point at a time from each of a plurality of pixels (contour pixels) through which the actual image contour passes. The contour position is extracted in sub-pixel units within each contour pixel. In the example of FIG. 3, the coordinates (x, y) within a pixel indicate the contour position. Further, the angle θ indicates an angle in the normal direction at each contour position of a contour approximated by fitting a plurality of contour positions with a predetermined function. The angle θ in the normal direction is defined, for example, as a clockwise angle with respect to the X axis. The obtained information on each actual image contour position (actual image outline data) is stored in the storage device 22. In this example, a case has been described in which a Sobel filter is used, but the contour extraction filter is not limited to a Sobel filter.

The image acquisition circuit 213 transmits the generated test image to the comparison circuit 214 and the storage device 22.

2.2 Reference image acquisition process Next, an example of the reference image acquisition process will be described. For example, the defect inspection device 1 acquires design data via the communication device 25 (step S21). The acquired design data is stored in the storage device 22, for example.

The expansion circuit 211 reads design data stored in the storage device 22. Then, the expansion circuit 211 executes expansion processing and expands (converts) the design data into, for example, 8-bit image data (expanded image) (step S22). Each pixel of the developed image has a value corresponding to the occupancy rate of that pixel by the figure of the design data. For example, in the case of 8-bit image data, the pixel value is 0 when the occupancy rate of the design figure is 0%, and the pixel value is 255 when the occupancy rate is 100%. The expansion circuit 211 transmits the expanded image to the reference image generation circuit 212 and the storage device 22.

The resizing processing circuit 215 executes resizing processing of the expanded image (step S23). Details of the resizing process will be described later.

Next, the corner rounding processing circuit 216 performs rounding processing on the corner portion of each resized pattern (step S24).

The reference image generation circuit 212 extracts the contour of the pattern from the expanded image that has been subjected to the resizing process and the corner rounding process (step S25), and generates a reference image (contour image). The reference image generation circuit 212 transmits the generated reference image to the comparison circuit 214 and the storage device 22.

2.3 Comparison Process Next, an example of the comparison process will be described. First, the comparison circuit 214 performs alignment using the inspection image and the reference image (step S31), and aligns the pattern in the inspection image with the pattern in the reference image. For example, a relative vector between each contour line position of the actual image (inspection image) and the corresponding contour line position of the reference image is obtained, and the average value thereof is taken as the alignment shift amount. At this time, the comparison circuit 214 calculates the amount of alignment shift of the test image with respect to the reference image.

Next, the comparison circuit 214 measures the amount of distortion of the inspection image (step S32) and calculates the distortion coefficient. For example, due to stage movement accuracy or distortion of the sample 300, a positional shift may occur between the coordinate information based on the design data and the coordinates of the pattern calculated from the captured image. The comparison circuit 214 measures the amount of distortion of the inspection image from, for example, the distribution of local alignment shift amounts within the surface of the sample 300, and calculates the distortion coefficient.

For example, the amount of distortion of coordinates (x, y) is expressed by the following equation, where dx and dy are the distortion amounts. Here, _a1 to _a4 and _b1 to _b4 are the distortion coefficients. Each distortion coefficient can be calculated by an optimization method such as the least squares method, with the relative vector (dx, dy) between each contour point of the inspection image and the corresponding contour point of the reference image.
dx(x,y) = _a1 + _a2x + _a3y + _a4xy
dy(x,y)= _b1 + _b2x + _b3y + _b4xy

Since the distortion amount at each contour point can be calculated using the calculated distortion coefficient, it is used together with the alignment shift amount in step S33, which will be described later, to calculate a relative vector between the inspection image and the reference image.

Next, the comparison circuit 214 compares the inspection image and the reference image (step S33). Comparison circuit 214 detects defects based on the comparison results. In other words, in the comparison process of comparing the inspection image and the reference image, the comparison circuit 214 uses the alignment shift amount to compare each contour line (actual image contour line) of the inspection image and the corresponding contour line ( (reference image outline). For example, the comparison circuit 214 determines the magnitude (distance) of the defect position deviation vector in consideration of the amount of alignment shift between each of the plurality of actual image contour positions and the corresponding reference contour position. If the threshold value is exceeded, it is determined to be a defect. The comparison result is output to the storage device 22 or the monitor 23.

After storing the results of the defect inspection in the memory device 22, the control circuit 21 may, for example, display the results on the display device 23, or output the results to an external device (e.g., a review device, etc.) via the communication device 25.

3. Resizing Process Next, an example of the resizing process will be described. Fig. 4 is a flowchart of the resizing process. As shown in Fig. 4, the resizing process circuit 215 executes steps S101 to S108. Each step will be described in detail.

[Step S101]
First, the resizing processing circuit 215 classifies each figure in the expanded image. That is, the resizing processing circuit 215 classifies each pixel including a pattern edge into a plurality of classes. The edge of a figure in a developed image can be considered to be at the boundary between a pixel with a gradation value of 0 and a pixel with a gradation value greater than 0. Here, a pixel with a gradation value greater than 0 and adjacent to a pixel with a gradation value of 0 is defined as an edge pixel.

Here, an example of class classification will be described. For example, in this embodiment, the pixel of interest is classified into four classes A to D, as shown in Figures 5 to 8. The number of classes and the definition of each class can be set arbitrarily. Figures 5 to 8 show the pixel of interest on the edge of a line pattern corresponding to classes A to D, respectively. Note that in the example of Figures 5 to 8, the pixel of interest is located on the edge on the right side of the pattern, but the left side and top and bottom edges of the pattern can be made to correspond by rotating Figures 5 to 8 by 90°, 180°, and 270°.

In the following description, the distance from the pixel of interest provided on the edge of a line pattern to the edge of another pattern facing each other is assumed to be L1. Let L2 be the distance from the pixel of interest to the opposing edge in the pattern. Let L3 be the distance from the edge where the pixel of interest is placed to one edge end (corner). Let L4 be the distance from the edge where the pixel of interest is placed to the other edge end (corner).

As shown in FIG. 5, in class A, in a line pattern extending in the This includes the case where the width) (L3+L4) is less than or equal to the threshold th2 ((L3+L4)≦th2). The threshold value th1 is a threshold value set in advance for the distance L1 from the pattern edge including the pixel of interest to the edge of another pattern facing the pattern edge. The threshold th2 is a threshold set in advance for the line width of the pattern including the pixel of interest.

The logical expression indicating class A is (L1≦th1)∧((L3+L4)≦th2). Here, “∧” indicates logical product.

As shown in FIG. 6, class B does not include class A, and in a line pattern extending in the X direction, the distance from the pixel of interest to the opposing edge (line length of the line pattern) (L2) is greater than or equal to the threshold th3. (L2≧th3) and the edge width (line width of the line pattern) (L3+L4) including the pixel of interest is less than or equal to the threshold th2 ((L3+L4)≦th2). The threshold th3 is a threshold set in advance for the line length of the pattern including the pixel of interest.

The logical expression indicating class B is (!A)∧(L2≧th3)∧((L3+L4)≦th2). Here, "!" indicates the negation of a logical operation.

As shown in FIG. 7, class C does not include classes A and B, and in a line pattern extending in the Y direction, the distance from the pixel of interest to the opposing edge (in this case, the line width of the line pattern) (L2 ) is less than or equal to a preset threshold th2 (L2≦th2).

The logical expression indicating class C is (!A)∧(!B)∧(L2≦th2).

As shown in Figure 8, class D does not include classes A to C. Therefore, the logical expression representing class D is (!A) ∧ (!B) ∧ (!C).

[Step S102]
Next, the resizing processing circuit 215 applies a resizing amount corresponding to the class to each edge (each pixel of the expanded image) according to the result of the classification of the figure. Then, the resizing processing circuit 215 creates a resizing amount map indicating the resizing amount of each pixel. FIG. 9 is an example of a resizing amount map.

As shown in FIG. 9, for example, a resizing amount based on class classification is given to each pixel including edges of four patterns P1 to P4. For example, a resizing amount of 100 is given to pixels classified as class A. A resizing amount of 75 is assigned to pixels classified as class B. A resizing amount of 50 is assigned to pixels classified as class C. A resizing amount of 25 is assigned to pixels classified as class D. In the following explanation, it is defined that when the resizing amount is 100, the edge moves by one pixel. Note that the amount of resizing for each class can be set arbitrarily.

For example, each pixel including the lower edge of pattern P1 faces pattern P2 and is classified into class A, and therefore is given a resize amount of 100. Each pixel including the upper edge of pattern P1 is classified into class B and is therefore given a resize amount of 75. Further, since each pixel including the left and right edges of the pattern P1 is classified into class C, a resize amount of 50 is given to the pixel.

For example, a pixel including an edge on the left side of pattern P2 that faces pattern P3 is classified into class A, and therefore is given a resize amount of 100. Other pixels including the left and right edges of pattern P2 are classified into class B and are therefore given a resize amount of 75. Further, since each pixel including the upper and lower edges of the pattern P2 is classified into class C, a resize amount of 50 is given to it. The same applies to pattern P3.

For example, pattern P4 indicates a pattern that is larger in size than patterns P1 to P3. Each pixel including the edge of pattern P4 is classified into class D, and therefore is given a resize amount of 25.

[Step S103]
Next, the resizing processing circuit 215 executes adjacent pixel MAX processing on the resizing amount map. Here, the adjacent pixel MAX processing will be explained using FIG. 10. FIG. 10 is a diagram illustrating an example of adjacent pixel MAX processing. Hereinafter, the coordinates of each pixel in the 3×3 resizing amount map are expressed as (x, y)=(1, 1) to (3, 3) from the bottom left of the page to the top right of the page.

As shown in FIG. 10, the adjacent pixel MAX process is a process in which, for a pixel with a resize amount set to 1 or more, if the resize amounts of the pixels adjacent above, below, left, and right are greater than the resize amount of the target pixel, the resize amount of the target pixel is replaced with the resize amount of the adjacent pixels. The adjacent pixel MAX process is repeated a preset number of times. More specifically, for example, in a 3×3 resize amount map, the resize amounts of the pixels at coordinates (1,1), (2,1), and (2,2) are set to 50, and the resize amount of the pixel at coordinates (2,3) is set to 100. In such a case, when the first adjacent pixel MAX process is applied, the resize amount of the pixel at coordinates (2,2) is replaced from 50 to 100. Next, when the second adjacent pixel MAX process is applied, the resize amount of the pixel at coordinates (2,1) is replaced with 100.

For example, if adjacent pixel MAX processing is applied once to the resizing amount map explained in FIG. , is expanding.

[Step S104]
Next, the resizing processing circuit 215 performs smoothing processing (filter processing) on the resizing amount. The smoothing process prevents steps from occurring in the shape after the resizing process due to discontinuous resizing amounts. Here, an example of smoothing processing for the amount of resizing will be described.

The resizing processing circuit 215 first performs the calculation of Equation (2) to which the weighting function (Gaussian function) of Equation (1) is applied, and calculates a predetermined range centered on the coordinates (x, y) of the pixel of interest. , the sum of the products of the n-th power of the resizing amount and the Gaussian function of the distance from the pixel of interest at the coordinates (x', y') of each pixel within the range is determined. Note that n is a sharpening coefficient for emphasizing edges, and can be set arbitrarily.

Here, g(ξ,η) is the weighting function, s is the σ (standard deviation) of the weighting function, and R is the amount of resizing before smoothing.

Next, the resizing processing circuit 215 performs the calculation of equation (3) to obtain the sum of the Gaussian functions of the pixels that have edges within the above range.

Here, E represents whether or not the pixel (x, y) includes an edge; if it includes an edge, E(x, y) = 1; if it does not include an edge, E(x, y) =0.

Next, the resizing circuit 215 performs the calculation of formula (4) to calculate the resizing amount R2 after the smoothing process. In formula (4), normalization is performed by dividing a(x, y) calculated by formula (2) by b(x, y) calculated by formula (3), thereby reducing the influence of the density of the number of edges within the range. Note that in formula (4), the larger of the calculation result and the resizing amount before the process is used, so that the resizing amount R2 after the smoothing process is not smaller than the resizing amount R before the smoothing process.

For example, when smoothing processing is applied to the resizing amount map described in FIG. become).

[Step S105]
Next, the resizing processing circuit 215 expands the resizing amount to adjacent pixels other than edges. Dilation processing is a process in which a resizing amount is given to a pixel with a resizing amount of 0 among pixels adjacent to a pixel for which a resizing amount of 1 or more is set, and the pixels to which the resizing amount is given are increased (dilated). This has the effect of preventing resizing processing from not being performed when the edge position is located between pixels. The expansion process is performed in the X direction and the Y direction, respectively. Note that the expansion process may be repeatedly executed a preset number of times.

First, an example of the expansion process in the X direction will be described with reference to Figs. 13 and 14. Fig. 13 is a conceptual diagram for explaining the expansion process in the X direction. Fig. 14 is a diagram showing an example of the expansion process in the X direction.

As shown in FIG. 13, for example, in a 3x3 resizing amount map centered on a pixel of interest with a resizing amount of 0, the resizing amount of the pixel at coordinates (1, 1) is set to a, and the resizing amount of the pixel at coordinates (1, 2) is Let b be the amount of resizing of a pixel, and let c be the amount of resizing of the pixel at coordinates (1, 3). For example, if the resizing amount of the pixel of interest after the expansion process in the X direction is d, then the maximum values of 0.7a, b, and 0.7c are set for d. Here, the pixels at coordinates (1, 1) and (1, 2) are arranged diagonally to the pixel of interest at coordinates (2, 2), so the distance from the pixel of interest in the X direction is Taking this into consideration, a and c are each set to 0.7 times.

As shown in FIG. 14, for example, the resizing amount of the pixel at coordinates (1, 1) is 50, the resizing amount of the pixel at coordinates (1, 2) is 150, and the resizing amount of the pixel at coordinates (1, 3) is 50. Expansion processing in the X direction is performed with the resizing amount being 120 and the resizing amounts of other pixels being 0. Focusing on the pixel at coordinates (2,2), the set value of the resizing amount for the pixel at coordinates (1,1) is 50×0.7=35. The set value of the resizing amount by the pixel at coordinates (1, 2) is 150. Further, the set value of the resizing amount by the pixel at the coordinates (1, 3) is 120×0.7=84. Therefore, the maximum value of 150 is assigned as the resizing amount for the pixel at coordinates (2, 2). Furthermore, focusing on the pixel at coordinates (2, 1), the set value of the resizing amount by the pixel at coordinates (1, 1) is 50. The set value of the resizing amount by the pixel at coordinates (1, 2) is 150×0.7=105. Therefore, the maximum value of 105 is assigned as the resizing amount for the pixel at coordinates (2, 1). Further, focusing on the pixel at coordinates (2, 3), the set value of the resizing amount by the pixel at coordinates (1, 2) is 150×0.7=105. The set value of the resizing amount by the pixel at coordinates (1, 3) is 120. Therefore, the maximum value of 120 is assigned as the resizing amount for the pixel at coordinates (2, 3). Here, for example, in calculating the resizing amount of the coordinates (2, 1), the resizing amount of the coordinates (1, 0), which is not shown, is referred to, but is omitted here.

Next, an example of expansion processing in the Y direction will be described using FIGS. 15 and 16. FIG. 15 is a conceptual diagram for explaining the expansion process in the Y direction. FIG. 16 is a diagram illustrating an example of expansion processing in the Y direction.

As shown in FIG. 15, for example, in a 3×3 resize amount map centered on a pixel of interest with a resize amount of 0, the resize amount of the pixel at coordinates (1,1) is α, the resize amount of the pixel at coordinates (2,1) is β, and the resize amount of the pixel at coordinates (3,1) is γ. For example, if the resize amount of the pixel of interest after expansion processing in the Y direction is δ, then, as with expansion processing in the X direction, δ is set to the maximum value of 0.7α, β, and 0.7γ.

As shown in FIG. 16, for example, the resizing amount of the pixel at coordinates (1,1) is 180, the resizing amount of the pixel at coordinates (2,1) is 30, and the resizing amount of the pixel at coordinates (3,1) is 180. Expansion processing in the Y direction is performed with the resizing amount being 80 and the resizing amounts of other pixels being 0. Focusing on the pixel at coordinates (2,2), the set value of the resizing amount for the pixel at coordinates (1,1) is 180×0.7=126. The set value of the resizing amount by the pixel at coordinates (2, 1) is 30. Further, the set value of the resizing amount by the pixel at the coordinates (3, 1) is 80×0.7=56. Therefore, the maximum value of 126 is assigned as the resizing amount for the pixel at coordinates (2, 2). Furthermore, focusing on the pixel at coordinates (1, 2), the set value of the resizing amount by the pixel at coordinates (1, 1) is 180. The set value of the resizing amount by the pixel at coordinates (2, 1) is 30×0.7=21. Therefore, the maximum value of 180 is assigned as the resizing amount for the pixel at coordinates (1, 2). Furthermore, focusing on the pixel at coordinates (3, 2), the set value of the resizing amount by the pixel at coordinates (2, 1) is 30×0.7=21. The set value of the resizing amount by the pixel at coordinates (3, 1) is 80. Therefore, the maximum value of 80 is assigned as the resizing amount for the pixel at coordinates (3, 2). As in the case of the X direction, reference to the resizing amount of coordinates (not shown) is omitted.

After the expansion process, the resizing amount in the X direction and the resizing amount in the Y direction are stored as two types of resizing amount maps in the X direction and the Y direction.

[Step S106]
Next, the resizing processing circuit 215 executes resizing processing pixel by pixel based on the resizing amount map. The resizing process in units of one pixel is, for example, a resizing process in which the upper limit of one resizing process is set to 100 when the resizing amount to move the edge position by one pixel is defined as 100. Here, resizing processing in units of one pixel will be explained using FIG. 17. FIG. 17 is a diagram illustrating an example of resizing processing. In FIG. 17, the pattern is a visualization of the pattern (figure) on the design data, and the developed image itself has a gradation value corresponding to the occupancy rate of the figure. Therefore, the total gradation value of the developed image is proportional to the area of the figure in that region. In FIG. 17, the resizing process is performed in the direction of expanding the figure, but when shrinking the figure, it is preferable to perform the resizing process after inverting the gradation of the figure in advance, and then inverting it again after the process.

As shown in FIG. 17, in a 3×3 expanded image centered on a pixel of interest including a pattern edge, the normal direction of the contour at the pixel of interest is determined using a Sobel filter. More specifically, the resizing processing circuit 215 performs a convolution operation on the expanded image with the X-direction and Y-direction kernels of the Sobel filter. Then, the resizing processing circuit 215 calculates the normal angle θ of the contour vector obtained by combining the calculated value for the X direction and the calculated value for the Y direction at the pixel of interest. For example, when the value in the X direction after Sobel filter processing is Fx, and the value in the Y direction is Fy, it is expressed by the formula θ=atan(Fy/Fx).

Next, the resizing processing circuit 215 determines that the sum of pixel values (gradation values) of the 3 x 3 expanded image (that is, the total area of 3 x 3 pixels) is the same, and that the contour vector and the boundary obtained by the Sobel filter are the same. The lines form a perpendicular half-plane.

Next, the resizing processing circuit 215 moves the half plane in the direction of the resizing amount at the pixel of interest by a resizing amount of at most one pixel or less (in this example, a resizing amount of 100 or less). That is, the resizing processing circuit 215 corrects the pixel value of the pixel of interest according to the resizing amount, and executes resizing processing in units of one pixel. The resizing amount has two components in the X direction and Y direction as a result of the expansion process described later. The amount of resizing in the figure is represented by a vector (Rx·cos θ, Ry·sin θ), where Rx is the amount of resizing in the X direction, and Ry is the amount of resizing in the Y direction.

[Step S107]
Next, the resizing processing circuit 215 subtracts the resizing amount for one pixel (100 in this example) from the resizing amount map.

[Step S108]
If the resizing amount of each pixel is less than or equal to a preset threshold (for example, 0) (step S108_Yes), the resizing process ends. On the other hand, if there remain pixels whose resizing amount is larger than a preset threshold (for example, 0) (step S108_No), the resizing processing circuit 215 proceeds to step S105 and executes the expansion process again. The resizing processing circuit 215 repeats the expansion process, the resizing process, and the subtracting process of the resizing amount (steps S105 to S108) until the resizing amount of each pixel becomes equal to or less than a preset threshold (for example, 0).

An example of repetition of expansion processing, resizing processing, and resizing amount subtraction processing will be described using FIG. 18. FIG. 18 is a diagram illustrating an example of repetition of expansion processing, resizing processing, and subtraction processing in the X direction of the resizing amount map. Although omitted here, the resizing amount map is also subjected to dilation processing as shown in FIG. 16 in the Y direction, and after the resizing processing in units of one pixel, subtraction processing is performed as described later.

As shown in FIG. 18, for example, in a 3×3 resizing amount map, the resizing amount of the pixel at coordinates (2, 1) is 250, the resizing amount of the pixel at coordinates (2, 2) is 200, Assume that the resizing amount of the pixel at coordinates (2, 3) is 100, and the resizing amount of other pixels is 0.

In this state, the resizing processing circuit 215 executes expansion processing (step S105). As a result, a resize amount of 250 is given to the pixels at coordinates (1, 1) and (3, 1). A resize amount of 200 is given to the pixels at coordinates (1, 2) and (3, 2). A resize amount of 140 is given to pixels at coordinates (1, 3) and (3, 3).

Next, the resizing processing circuit 215 executes a pixel-by-pixel resizing process with a maximum resizing amount of 100 (step S106) and a process of subtracting the resizing amount (step S107). As a result, the resizing amount of pixels at coordinates (1,1), (2,1), and (3,1) is 150. The amount of resizing of pixels at coordinates (1, 2), (2, 2), and (3, 2) is 100. The amount of resizing for pixels at coordinates (1, 3) and (3, 3) is 40. The amount of resizing of the pixel at coordinates (2, 3) becomes 0.

In this state, since there remain pixels with a resizing amount of 1 or more (step S108_No), the resizing processing circuit 215 repeats the process and executes the dilation process again (step S105). As a result, a resize amount of 70 is given to the pixel at coordinates (2, 3).

Next, the resizing processing circuit 215 executes a pixel-by-pixel resizing process with a maximum resizing amount of 100 (step S106) and a process of subtracting the resizing amount (step S107). As a result, the resizing amount of pixels at coordinates (1,1), (2,1), and (3,1) is 50. The amount of resizing for pixels at other coordinates is 0.

The resizing processing circuit 215 repeats the process until the resizing amount of each pixel becomes 0, since there are pixels whose resizing amount is 1 or more (step S108_No).

4. Extraction of Contour Points Next, an example of extraction of contour points in a developed image will be described using FIG. 19.
FIG. 19 is a diagram illustrating an example of extraction of contour points in a developed image. In FIG. 19, the pattern is a visualization of the pattern (figure) on the design data, and the developed image itself has a gradation value corresponding to the occupancy rate of the figure. Therefore, the total gradation value of the expanded image is proportional to the area of the figure in that region.

As shown in FIG. 19, in the same manner as in FIG. 17, a contour vector of a pixel of interest is found using a Sobel filter, and a corresponding half-plane is formed. A contour point is then set at the midpoint in the X and Y directions on the boundary line of the half-plane. The contour position is set in sub-pixel units within each contour pixel. As with the example in FIG. 3, it is preferable to show it as coordinates (x, y) within the pixel.

5. Effects of this Embodiment With the configuration of this embodiment, the defect inspection apparatus classifies patterns based on the size, shape, and arrangement of the patterns when generating a reference image, and applies different resizing amounts for each class. Can be set. Therefore, the defect inspection apparatus can generate reference image data with different resizing amounts based on the size and shape of the pattern. As a result, it is possible to reduce the deviation between the inspection image and the reference image in which the dimension shift has occurred. Therefore, it is possible to reduce the extraction of false defects during defect inspection, and improve the reliability of defect inspection.

6. Modifications, etc. In the above-described embodiment, a case has been described in which a Gaussian function is used for the smoothing process of the resizing amount, but other smoothing filters may be used. Further, although the case has been described in which a Sobel filter is used to calculate the contour vector, other contour extraction filters may be used.

In the above-described embodiment, a case has been described in which a reference image is generated in a defect inspection apparatus, but the method for generating a reference image is not limited to the defect inspection apparatus. The present invention may also be applied to other devices such as devices that generate reference images based on data, such as measurement devices.

In the embodiments described above, machine learning or the like may be used to classify patterns. That is, AI (artificial intelligence) may learn patterns and set calculation conditions for class classification.

Note that the present invention is not limited to the above-described embodiments, and can be variously modified at the implementation stage without departing from the gist thereof. Moreover, each embodiment may be implemented in combination as appropriate, and in that case, the combined effect can be obtained. Furthermore, the embodiments described above include various inventions, and various inventions can be extracted by combinations selected from the plurality of constituent features disclosed. For example, if a problem can be solved and an effect can be obtained even if some constituent features are deleted from all the constituent features shown in the embodiment, the configuration from which these constituent features are deleted can be extracted as an invention.

DESCRIPTION OF SYMBOLS 1... Defect inspection device, 10... Imaging mechanism, 11... Sample chamber, 12... Lens barrel, 13... Stage, 14... Stage drive mechanism, 15... Detector, 16... Electron gun, 17... Electron optical system, 20... Control Mechanism, 21... Control circuit, 22... Storage device, 23... Display device, 24... Input device, 25... Communication device, 101, 102... Focusing lens, 103, 104... Scanning coil, 105... Objective lens, 211... Deployment circuit , 212... Reference image generation circuit, 213... Image acquisition circuit, 214... Comparison circuit, 215... Resize processing circuit, 216... Corner rounding processing circuit, 221... Design data, 222... Parameter information, 223... Inspection data, 224... Defect Inspection program, 300...sample

Claims (8)

a sample imaging mechanism;
an image acquisition circuit that generates an inspection image based on image data of the sample captured by the imaging mechanism;
an expansion circuit that generates an expansion image from design data;
Classify the pattern from the developed image, apply the resizing amount to the edge of the pattern based on the resizing amount set for each classified class, and create a resizing amount map showing the resizing amount of each pixel. create, execute adjacent pixel maximization processing of the resizing amount map, execute smoothing processing of the resizing amount map after the adjacent pixel maximization processing, and execute expansion processing of the resizing amount map after the smoothing processing. and a resizing processing circuit that executes a resizing process to move the position of the edge of the pattern pixel by pixel based on the resizing amount map after the expansion process, and using the developed image after the resizing process. a reference image generation circuit that generates a reference image;
a comparison circuit that compares the inspection image and the reference image ,
The resizing processing circuit is
In the adjacent pixel maximization process, the resizing amount of a second pixel adjacent to any of the first pixels for which a first predetermined value or more is set as the resizing amount along the edge of the pattern. is larger than the resizing amount of the first pixel, replacing the resizing amount of the first pixel with the resizing amount of the second pixel,
In the dilation process, among adjacent pixels other than the edge of the pattern on the first pixel for which the resizing amount is set to a second predetermined value or more, the resizing amount is a third predetermined value. giving a certain third pixel a resizing amount based on the resizing amount of the first pixel;
Defect inspection equipment. The resizing processing circuit is
In the classification of the pattern, a first distance from at least a pixel of interest provided on a first edge of the pattern to another pattern facing the pattern, and a distance from the pixel of interest to another pattern facing the pixel of interest in the pattern; a second distance to a second edge; a third distance from the first edge where the pixel of interest is placed to a first end of the first edge; and a third distance from the first edge where the pixel of interest is placed to the first edge. performing the classification of the pattern based on a fourth distance of the first edge to the second end;
In the smoothing process, the pixels within a preset range centered on the pixel of interest are resized based on a Gaussian function based on the distance from the pixel of interest and a weight of a power of the resizing amount of the pixel. filter amount,
The defect inspection device according to claim 1. The defect inspection apparatus according to claim 2 , wherein the resizing processing circuit executes a process of subtracting the resizing amount applied to the resizing process from the resizing amount map after the resizing process for each pixel . The resizing processing circuit repeats the expansion processing, the resizing processing, and the subtraction processing until the resizing amount in the resizing amount map after the subtraction processing becomes equal to or less than a preset threshold.
The defect inspection device according to claim 3. a step of imaging the sample to generate an inspection image;
a step of generating a developed image from design data;
a step of classifying patterns from the developed image;
setting a resizing amount for each classified class, applying the resizing amount to the edge of the pattern, and creating a resizing amount map showing the resizing amount of each pixel ;
a step of executing adjacent pixel maximization processing of the resizing amount map;
performing smoothing processing on the resizing amount map after the adjacent pixel maximization processing;
performing expansion processing of the resizing amount map after the smoothing processing;
executing a resizing process of moving the position of the edge of the pattern pixel by pixel based on the resizing amount map after the expansion process ;
generating a reference image based on the developed image after the resizing process;
and performing an inspection by comparing the inspection image and the reference image ,
In the step of executing the adjacent pixel maximization process, a second pixel adjacent to any of the first pixels for which a first predetermined value or more is set as the resizing amount along the edge of the pattern; If the resizing amount is larger than the resizing amount of the first pixel, replacing the resizing amount of the first pixel with the resizing amount of the second pixel,
In the step of executing the dilation process, among adjacent pixels other than the edge of the pattern on the first pixel for which the resizing amount is set to a second predetermined value or more, the resizing amount is a third predetermined value. giving a resizing amount based on the resizing amount of the first pixel to a third pixel that is a predetermined value;
Defect inspection method. In the step of classifying the pattern, a first distance from at least a pixel of interest provided on a first edge of the pattern to another pattern facing the pattern, and a distance from the pixel of interest to the pixel of interest within the pattern. a third distance from the first edge where the pixel of interest is placed to a first end of the first edge; and the first distance where the pixel of interest is placed. performing the classification of the pattern based on a fourth distance from an edge to a second end of the first edge;
In the step of executing the smoothing process, the smoothing process is performed on pixels within a preset range centered on the pixel of interest based on a Gaussian function based on the distance from the pixel of interest and a weight of a power of the resizing amount of the pixel. the resizing amount is filtered ;
The defect inspection method according to claim 5. The method further comprises the step of subtracting the resize amount applied to the resize processing from the resize amount map after the resize processing for each pixel.
The defect inspection method according to claim 6. repeating the expansion process, the resizing process, and the subtraction until the resizing amount in the resizing amount map after the subtraction becomes equal to or less than a preset threshold;
The defect inspection method according to claim 7.