Identification of arrays in semiconductor samples

Abstract

To provide identification of an array in a semiconductor specimen.SOLUTION: There are provided a method and a system configured to obtain an image of a semiconductor specimen including one or more arrays, each including repetitive structural elements, and one or more regions, each region at least partially surrounding a corresponding array and including features different from the repetitive structural elements. A PMC is configured to, during run-time scanning of the semiconductor specimen, perform a correlation analysis between pixel intensity of the image and pixel intensity of a reference image informative of at least one of the repetitive structural elements, to obtain a correlation matrix, use the correlation matrix to distinguish between one or more first areas of the image corresponding to the one or more arrays and one or more second areas of the image corresponding the one or more regions, and output data informative of the one or more first areas of the image.SELECTED DRAWING: Figure 2

Description

[Technical Field] 【0001】 The subject matter of this disclosure generally relates to the field of sample testing, and more specifically to the automation of sample testing. [Background technology] 【0002】 The current demands for high density and performance associated with the ultra-large-scale integration of manufactured devices necessitate submicron feature areas, increased transistor and circuit speeds, and improved reliability. Such demands require the formation of highly accurate and uniform device feature areas, and consequently, careful monitoring of the manufacturing process, including automated testing of the device while it is still in the form of a semiconductor wafer. 【0003】 Testing processes are used at various steps in semiconductor manufacturing to detect and classify defects in samples. The effectiveness of testing can be improved by automating processes such as automated defect classification (ADC) and automated defect review (ADR). [Overview of the project] 【0004】 According to a particular aspect of the subject matter of the present disclosure, a system is provided for testing a semiconductor sample, the system comprising a processor and memory circuit (PMC) configured to acquire an image of the semiconductor sample comprising one or more arrays, each comprising a repeating structural element, and one or more regions, each region comprising a feature distinct from the repeating structural element, at least partially surrounding a corresponding array, wherein the PMC is configured to perform a correlation analysis between the pixel intensity of the image and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements in order to acquire a correlation matrix, using the correlation matrix to distinguish one or more first regions of the image corresponding to one or more arrays from one or more second regions of the image corresponding to one or more regions, and outputting data that provides information on one or more first regions of the image. 【0005】 According to some embodiments, the system is configured to determine subregions of an image corresponding to correlation matrix values ​​that satisfy an amplitude criterion, cluster the subregions into one or more clusters based on data that provides distance information between repeating structural elements in an array, and determine one or more first regions based on at least one or more clusters. 【0006】 According to some embodiments, one or more arrays are separated from one or more regions by one or more boundaries, and the system is configured to estimate one or more first regions of an image that include only one or more arrays up to the boundaries. 【0007】 According to some embodiments, the system is configured to apply image processing to a reference image, and the image processing attenuates repeating patterns in the reference image. 【0008】 According to some embodiments, the system is configured to cluster subregions into one or more first clusters based on data that provides distance information between repeating structural elements in an array along a first axis, cluster subregions into one or more second clusters based on data that provides distance information between repeating structural elements in an array along a second axis, and to use the first and second clusters to distinguish one or more first regions of an image corresponding to one or more arrays from one or more second regions of an image corresponding to one or more regions. 【0009】 According to some embodiments, the system is configured to determine a polygon enclosing one or more clusters for each cluster and to output the polygon as a first region of the image. 【0010】 According to some embodiments, the system is configured to select only clusters where several sub-regions meet a threshold. 【0011】 According to some embodiments, the system is configured to acquire data that provides amplitude-reference information during the setup phase prior to runtime testing of a semiconductor sample. 【0012】 According to some embodiments, the system is configured to perform a correlation analysis between the pixel intensity of one or more first regions of an image and the pixel intensity of a second reference image that provides information on at least one of the repeating structural elements in order to obtain a second correlation matrix; to determine subregions of one or more first regions of the image that correspond to values ​​of the second correlation matrix that satisfy an amplitude criterion; to determine a distortion map between one or more first regions of the image and the array based at least on the location of the subregions in one or more first regions of the image and data that provides information on the intended location of the repeating structural elements in the array; and to generate a corrected image based on the distortion map. 【0013】 According to some embodiments, the system is configured to generate a corrected image such that data providing information on the locations of subregions in the corrected image and the intended locations of repeating structural elements in the array satisfy proximity criteria. 【0014】 According to some embodiments, this system provides a distortion DF between the position of subregions of one or more first regions of an image and data that provides information about the intended positions of repeating structural elements in an array. central To decide, at least DF central It is configured to determine a distortion map between one or more first regions of an image and an array of semiconductor samples, based on an interpolation method applied to the image. 【0015】 According to some embodiments, the system is configured to acquire a reference image that provides information on at least one of the repeating structural elements, and to select only a subset of the reference image as a second reference image. 【0016】 In other aspects of the subject matter of this disclosure, a method is provided for testing a semiconductor sample, the method comprising: acquiring an image of the semiconductor sample by a processor and memory circuit (PMC) of one or more arrays, each containing repeating structural elements, and one or more regions, each region, at least partially surrounding a corresponding array and containing feature portions distinct from the repeating structural elements; performing a correlation analysis between the pixel intensity of the image and the pixel intensity of a reference image providing information on at least one of the repeating structural elements in order to acquire a correlation matrix during a runtime scan of the semiconductor sample; using the correlation matrix to distinguish one or more first regions of the image corresponding to one or more arrays from one or more second regions of the image corresponding to one or more regions; and outputting data providing information on one or more first regions of the image. 【0017】 According to some embodiments, the method includes determining subregions of an image corresponding to correlation matrix values ​​that satisfy an amplitude criterion, clustering the subregions into one or more clusters based on data that provides distance information between repeating structural elements in an array, and determining one or more first regions based on at least one or more clusters. 【0018】 According to some embodiments, one or more arrays are separated from one or more regions by one or more boundaries, and the method includes estimating one or more first regions of an image that includes only one or more arrays up to the boundary, and excluding one or more second regions corresponding to one or more regions. 【0019】 According to some embodiments, the method includes clustering subregions into one or more first clusters based on data that provides distance information between repeating structural elements in an array along a first axis; clustering subregions into one or more second clusters based on data that provides distance information between repeating structural elements in an array along a second axis; and using the first and second clusters to distinguish one or more first regions of an image corresponding to one or more arrays from one or more second regions of an image corresponding to one or more regions. 【0020】 According to some embodiments, this method involves selecting only clusters where several sub-regions satisfy a threshold. 【0021】 According to some embodiments, the method includes: performing a correlation analysis between the pixel intensity of one or more first regions of an image and the pixel intensity of a second reference image that provides information on at least one of the repeating structural elements in order to obtain a second correlation matrix; determining subregions of one or more first regions of an image that correspond to values ​​of the second correlation matrix that satisfy the intensity criteria; determining a distortion map between one or more first regions of an image and an array, based at least on the location of the subregions in one or more first regions of an image and data that provides information on the intended location of the repeating structural elements in an array; and generating a corrected image based on the distortion map. 【0022】 According to some embodiments, this method provides a distortion DF between the position of subregions of one or more first regions of an image and data that provides information about the intended positions of repeating structural elements in an array. central To decide, and at least DF central This includes determining a distortion map between one or more first regions of an image and an array of semiconductor samples, based on an interpolation method applied to the image. 【0023】 According to other aspects of the subject matter of this disclosure, a non - transient computer - readable medium including instructions is provided, where when the instructions are executed by a PMC, the PMC is caused to perform operations as described above. 【0024】 According to other aspects of the subject matter of this disclosure, a system for testing a semiconductor sample is provided, the system comprising a processor and a memory circuit (PMC) configured to obtain an image of a semiconductor sample including one or more arrays each including repetitive structural elements and one or more regions each surrounding at least partially a corresponding array and including features different from the repetitive structural elements, and to obtain data D providing information on pixel intensities of at least one of the one or more arrays and the one or more regions. threshold The PMC is configured to determine data D representing pixel intensities along a plurality of axes of the image during runtime scanning of the semiconductor sample, X D Y and to use D, D, and D to distinguish one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more regions, and is configured to output data providing information on one or more first regions of the image. X ,D Y , and D threshold 【0025】 In some embodiments, the system is configured to determine data D representing pixel intensities along each of a plurality of rows of the image, X to determine data D representing pixel intensities along each of a plurality of columns of the image, Y and to use D, D, and D to distinguish one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more regions, and is configured to output data providing information on one or more first regions of the image. X ,D Y , and D threshold 【0026】 According to some embodiments, each of one or more arrays includes structural elements that cannot be distinguished by visual inspection of the image. 【0027】 According to some embodiments, this system uses data D representing the pixel intensity along each of multiple rows of an image. X Decide, D X A subset of images S that includes rows of images where the first threshold is exceeded. L Select subset S L Data D represents the pixel intensity along each of the multiple columns. Y,SL Decide, D Y,L S exceeds the second threshold. L Subset C of the column SL Determine C SL It is configured to determine one or more first areas based on at least the following. 【0028】 According to some embodiments, this system uses data D representing the pixel intensity along each of multiple columns of an image. Y Decide, D Y A subset of images S that includes a column of images where the first threshold is exceeded. C Select subset S C Data D represents the pixel intensity along each of the multiple rows. X,SC Decide, D X,SC S exceeds the second threshold. C L, a subset of the row SC Determine L SC It is configured to determine one or more first areas based on at least the following. 【0029】 According to some embodiments, the first threshold is stricter than the second threshold. According to some embodiments, the first threshold is stricter than the third threshold. 【0030】 According to some embodiments, a corresponding method (including operations as described above with reference to a system) and a non-temporary computer-readable medium containing instructions, wherein when an instruction is executed by a PMC, the PMC causes the PMC to perform a corresponding operation. 【0031】 According to several embodiments, the proposed solution enables the distinction in an image of a semiconductor sample between an array containing repeating structural elements and a surrounding region containing features distinct from the repeating structural elements. According to several embodiments, the proposed solution is efficient and functions during runtime scanning of the semiconductor sample. According to several embodiments, accurate identification of the array is achieved, thereby enabling the extraction of the array up to the array boundary separating the array from the surrounding region. According to several embodiments, the proposed solution enables the correction of distortion present in the image of the array. In particular, efficient and accurate correction is possible. 【0032】 To understand this disclosure and how it is actually put into practice, embodiments are now described with reference to the accompanying drawings, merely as non-limiting examples. [Brief explanation of the drawing] 【0033】 [Figure 1] This is a generalized block diagram of a test system according to a particular embodiment of the subject matter of this disclosure. [Figure 2] This is a generalized flowchart of a method for identifying arrays containing repeating structural elements in images of a sample. [Figure 2A] This figure shows a non-restrictive example of an image that includes an array with repeating structural elements and the surrounding region. [Figure 3] Figure 2 is a generalized flowchart of possible embodiments of the operation of the method. [Figure 3A] This figure shows a non-restrictive example of a correlation matrix obtained using the method shown in Figure 2. [Figure 3B] This figure shows a non-restrictive example of an array containing repeating structural elements. [Figure 3C]This figure shows another non-restrictive example of a correlation matrix obtained using the method in Figure 2. [Figure 4] This is a generalized flowchart of the method for processing the reference image used in the method shown in Figure 2. [Figure 4A-4B] This figure shows a non-restrictive example of using the method in Figure 4. [Figure 5] This is a generalized flowchart of another method for identifying arrays containing repeating structural elements in images of a sample. [Figure 5A] This figure shows a non-restrictive example of the application of the method in Figure 5. [Figure 6] This is a generalized flowchart of another embodiment of a method for identifying arrays containing repeating structural elements in an image of a sample. [Figure 6A-6C] This figure shows a non-restrictive example of the application of the method in Figure 6. [Figure 7] This figure shows a non-restrictive example of a distorted image of an array containing repeating structural elements. [Figure 8] This is a generalized flowchart of a method for correcting distortion in the image shown in Figure 7. [Figures 8A-8C] This figure shows a non-restrictive example of the application of the method in Figure 8. [Figure 9] This is a generalized flowchart of the method for processing the reference image used in the method shown in Figure 8. [Figure 9A] This figure shows a non-restrictive example of the application of the method in Figure 9. [Modes for carrying out the invention] 【0034】 The following detailed description includes numerous specific details to provide a complete understanding of the disclosure. However, those skilled in the art will understand that the subject matter of this disclosure can be practiced without these specific details. In other examples, well-known methods, procedures, components, and circuits are not described in detail so as not to obscure the subject matter of this disclosure. 【0035】 Unless otherwise specified, as will be apparent from the following discussion, any discussion throughout this specification using terms such as “process,” “acquire,” “select,” “determine,” “generate,” “output,” “use,” and “execute” should be understood to refer to computer actions and / or processes that manipulate and / or transform data into other data, where such data is represented as a physical quantity such as electrons, and / or represents a physical object. The term “computer” should be interpreted broadly to include, in non-limiting examples, any hardware-based electronic device of any kind with data processing capabilities, including System 103 and its respective parts disclosed in this application. 【0036】 As used herein, the terms “non-temporary memory” and “non-temporary storage medium” should be interpreted broadly to encompass any volatile or non-volatile computer memory suitable for the subject matter of this disclosure. 【0037】 As used herein, the term “sample” should be interpreted broadly to encompass wafers, masks, and other structures, combinations thereof, and / or parts thereof, of any kind used to produce semiconductor integrated circuits, magnetic heads, flat panel displays, and other semiconductor products. 【0038】 As used herein, the term “testing” should be interpreted broadly to encompass all kinds of metrology-related operations, as well as operations related to the detection and / or classification of defects in specimens during production. Testing is performed during or after the production of the specimen to be tested, by using non-destructive testing tools. In non-limiting examples, the testing process may include runtime scanning (in a single or multiple scans), sampling, review, measurement, classification, and / or other operations performed on the specimen or a portion thereof using the same or different inspection tools. Similarly, testing may be performed before the production of the specimen to be tested, and may include, for example, generating test strategies and / or other setting operations. Unless otherwise specified, it should be noted that the term “testing” or its derivatives as used herein is not limited in terms of the resolution or size of the inspection area. A wide variety of non-destructive testing tools include, in non-limiting examples, scanning electron microscopes, atomic force microscopes, optical inspection tools, and the like. 【0039】 As a non-limiting example, runtime testing can utilize a two-phase procedure, for example, inspection of the sample followed by a review of the sampled locations of potential defects. During the first phase, the surface of the sample is inspected quickly and at a relatively low resolution. In the first phase, a defect map is created to indicate suspicious locations on the sample with a high probability of defects. During the second phase, at least some of the suspicious locations are analyzed more thoroughly at a relatively high resolution. Both phases may be performed with the same inspection tool, or they may be performed with different inspection tools. 【0040】 As used herein, the term “defect” should be interpreted broadly to include any kind of abnormal or undesirable feature formed on or within a specimen. 【0041】 As used herein, the term “design data” should be interpreted broadly to encompass any data that demonstrates the hierarchical physical design (layout) of a sample. Design data may be provided by the respective designer and / or derived from the physical design (e.g., by complex simulations, simple geometric and Boolean operations, etc.). Design data may be provided in various formats, such as GDSII format, OASIS format, etc., as non-limiting examples. Design data may be presented in vector format, grayscale intensity image format, or otherwise. 【0042】 Unless otherwise noted, certain features of the subject matter of this disclosure described in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features of the subject matter of this disclosure described in the context of a single embodiment may also be provided separately or in any suitable sub-combination. The following detailed description includes numerous specific details to provide a complete understanding of the methods and apparatus. 【0043】 With this in mind, we will now focus on Figure 1, which shows a functional block diagram of a test system according to a particular embodiment of the subject matter of this disclosure. The test system 100 shown in Figure 1 can be used to test a sample (e.g., a wafer and / or a portion thereof) as part of a sample manufacturing process. The illustrated test system 100 includes a computer-based system 103 that can automatically determine metrology-related information and / or defect-related information using images acquired during sample manufacturing. The system 103 can be operably connected to one or more low-resolution test tools 101 and / or one or more high-resolution test tools 102 and / or other test tools. The test tools are configured to capture images and / or review captured images and / or enable or perform measurements related to captured images. The system 103 can further be operably connected to a CAD server 110 and a data repository 109. 【0044】 System 103 includes a processor and memory circuit (PMC) 104 operably connected to a hardware-based input interface 105 and a hardware-based output interface 106. The PMC 104 is configured to perform all the processing necessary to operate System 103 (see, for example, the methods described in Figures 2 to 5, 6, 8, and 9, which can be performed at least partially by System 103), and includes a processor (not shown separately) and memory (not shown separately). The processor of the PMC 104 can be configured to execute several functional modules according to computer-readable instructions implemented in non-temporary computer-readable memory contained within the PMC. Such functional modules are referred to below as contained within the PMC. The functional modules contained within the PMC 104 include a deep neural network (DNN) 112. The DNN 112 is configured to enable data processing using machine learning algorithms to output application-relevant data based on images of a sample. 【0045】 As a non-restrictive example, the layers of DNN112 can be organized according to a convolutional neural network (CNN) architecture, a recurrent neural network architecture, a generative adversarial network (GAN) architecture, or other methods. Optionally, at least some of the layers can be organized into multiple DNN subnetworks. Each layer of a DNN can contain a number of fundamental computational elements (CEs), commonly referred to in the art as dimensions, neurons, or nodes. 【0046】 In general, the computational elements of a given layer can be connected to the CEs of preceding and / or succeeding layers. Each connection between the CEs of the preceding and succeeding layers is associated with a weighted value. A given CE can receive input from the CE of the previous layer through each connection, and each given connection is associated with a weighted value that can be applied to the input of the given connection. The weighted values ​​can determine the relative strength of the connections, and therefore the relative influence of each input on the output of the given CE. A given CE can be configured to compute an activation value (e.g., a weighted sum of inputs) and then derive an output by applying an activation function to the computed activation. The activation function can be, for example, an identity function, a deterministic function (e.g., linear, sigmoid, threshold, etc.), a stochastic function, or other appropriate function. The output from a given CE can be sent to the CE of the succeeding layer through each connection. Similarly, as described above, each connection at the output of a CE can be associated with a weighted value that can be applied to the output of the CE before it is received as an input to the CE of the succeeding layer. In addition to weighted values, there may be thresholds (including limiting functions) associated with connections and CEs. 【0047】 The weights and / or thresholds for DNN112 can be initially selected before training, and then iteratively adjusted or modified during training to achieve an optimal set of weights and / or thresholds in the trained DNN. After each iteration, the difference (also called the loss function) between the actual output produced by DNN112 and the target output associated with each training dataset can be determined. This difference is sometimes called the error value. Training can be considered complete when the cost function or loss function indicating the error value is less than a predetermined value, or when a limited change in performance between iterations is achieved. Optionally, at least some (if any) of the DNN subnetworks may be trained separately before training the entire DNN. 【0048】 System 103 is configured to receive input data via input interface 105. Input data may include data produced by the test tool (and / or its derivatives and / or associated metadata), and / or data produced and / or stored in one or more data repositories 109, and / or in the CAD server 110 and / or other associated data repositories. It should be noted that input data may include images (e.g., capture images, images derived from capture images, simulated images, composite images, etc.) and associated numerical data (e.g., metadata, handcrafted attributes, etc.). It should be further noted that image data may include data related to the layer of interest, and / or data related to one or more other layers of the sample. 【0049】 System 103 is further configured to process at least a portion of the received input data and send the results (or a portion thereof) via the output interface 106 to a storage system 107, a test tool, a computer-based graphical user interface (GUI) 108 for rendering the results, and / or an external system (e.g., a FAB yield management system (YMS)). The GUI 108 can be further configured to allow user-specified input related to the operating system 103. 【0050】 As a non-limiting example, the sample may be tested with one or more low-resolution testing machines 101 (e.g., optical inspection systems, low-resolution SEMs, etc.). Result data providing information on the low-resolution image of the sample (hereinafter referred to as low-resolution image data 121) can be sent to system 103 (directly or via one or more intermediate systems). Alternatively or additionally, the sample may be tested with a high-resolution machine 102 (e.g., a subset of potential defect locations selected for review can be reviewed by a scanning electron microscope (SEM) or atomic force microscope (AFM)). Result data providing information on the high-resolution image of the sample (hereinafter referred to as high-resolution image data 122) can be sent to system 103 (directly or via one or more intermediate systems). 【0051】 It should be noted that images of desired locations on a sample can be captured at different resolutions. As a non-limiting example, a so-called "defect image" of a desired location can be used to distinguish between defects and false alarms, while a so-called "class image" of the desired location can be acquired at a higher resolution and used for defect classification. In some embodiments, multiple images of the same location (at the same or different resolutions) may include several images registered between them (e.g., multiple images captured from a given location and one or more reference images corresponding to a given location). 【0052】 It should be noted that image data may be received and processed along with its associated metadata (e.g., pixel size, text description of defect type, parameters of the image capture process, etc.). 【0053】 After processing the input data (e.g., low-resolution image data and / or high-resolution image data, optionally along with other data such as design data, synthesis data, etc.), system 103 can send the results (e.g., instruction-related data 123 and / or 124) to one of the test tools, store the results (e.g., defect attributes, defect classification, etc.) in storage system 107, render the results via GUI 108, and / or send them to an external system (e.g., YMS). 【0054】 Those skilled in the art will understand that the teachings of the subject matter of this disclosure are not constrained by the system shown in Figure 1, and equivalent and / or modified functions may be integrated or separated in other ways, and the software may be appropriately combined with the firmware and / or hardware. 【0055】 It should also be noted that, without limiting the scope of this disclosure, the test tool can be implemented as various types of inspection machines, such as optical imaging machines and electron beam inspection machines. In some cases, the same test tool can provide low-resolution and high-resolution image data. In some cases, at least one test tool can have metrological functions. 【0056】 It should be noted that the test system shown in Figure 1 can be implemented in a distributed computing environment, and the functional modules shown in Figure 1 can be distributed across several local and / or remote devices and linked through a communication network. Furthermore, it should be noted that in other embodiments, at least several test tools 101 and / or 102, a data repository 109, a storage system 107, and / or a GUI 108 may be located outside the test system 100 and be able to communicate data with system 103 via input interface 105 and output interface 106. System 103 can be implemented as a standalone computer that can be used in conjunction with the test tools. Alternatively, each function of the system can be at least partially integrated into one or more test tools. 【0057】 Next, we consider Figure 2. The method includes acquiring an image 250 of the sample (operation 200). According to some embodiments, the image 250 is acquired by a test tool (e.g., test tool 101) during a runtime scan of the sample. The sample comprises one or more arrays 260. The arrays 260 contain iterative structural elements (represented as reference 261 in one of the arrays). The iterative structural elements include, for example, memory cells (SRAM, DRAM, FRAM, flash memory, etc.), programmable logic cells, etc. These examples are not limiting. In general, the iterative structural elements are arranged in each array according to an iterative pattern or grid. For example, the distance between two adjacent iterative structural elements (by the horizontal and vertical axes) is constant or at least substantially constant between different arrays. 【0058】 The sample comprises one or more regions 265. Each region 265 encloses at least partially a corresponding array 260. The regions 265 do not contain any repeating structural elements present within the array 260. In the non-limiting example of Figure 2A, the sample includes vertical and horizontal regions 265 enclosing the array 260. The regions 265 may, for example, correspond to stitches. Each region 265 contains features distinct from the repeating structural elements 260. In some embodiments, the regions 265 may include non-repeatable features and / or repeatable features distinct from the repeating structural elements 260. An example of a non-repeatable feature is, for example, logic. However, this is not limiting. 【0059】 This method further includes performing a correlation analysis (operation 210) between the pixel intensity of image 250 and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements. The reference image may include, for example, an image of one of the repeating structural elements. The reference image is also called a “golden cell”. According to some embodiments, the reference image is generated based on design data. According to some embodiments, the reference image is obtained from images of structural elements that are known to be defect-free (for example, from a previous analysis). According to some embodiments, the reference image is obtained during a setup phase before runtime testing of the sample. In the setup phase, time and processing constraints are not as strict, and therefore it is possible to obtain an image of one of the repeating structural elements that make up the reference image. 【0060】 The output of the correlation analysis performed in 210 may include a correlation matrix containing multiple values. Each value is associated with a subregion in image 250 and indicates the level of correlation between the pixel intensity of the subregion and the pixel intensity of the reference image. 【0061】 This method may further include using a correlation matrix (operation 220) to distinguish one or more first regions of an image 250 corresponding to one or more arrays 260 from one or more second regions of an image corresponding to one or more regions 265. Figure 2A shows an example of one or more first regions 270 and one or more second regions 275. In some embodiments, all regions of an image not identified as belonging to one or more first regions 270 are considered to be part of one or more second regions 275. 【0062】 The method further includes outputting data 230 that provides information about one or more first regions 270 of image 250. This may include, for example, outputting the locations of one or more first regions 270 in image 250, and / or outputting a selection of image 250 that includes only one or more first regions 270. According to some embodiments, the method may include outputting the locations of one or more second regions 275, and / or outputting a selection of image 250 that includes only one or more second regions 275. 【0063】 According to some embodiments, at least operations 210, 220, and 230 are performed during the runtime scanning of the sample. In other words, the method for identifying the array in the image is efficient and therefore can be performed during the runtime phase. 【0064】 According to some embodiments, the identification of one or more first regions 270 in an image is used by a PMC configured to determine, for example, data representing defects in an array (e.g., location of the defect, class of the defect, etc.) during a runtime scan of the sample. In particular, the PMC may implement an algorithm for detecting defects that is specifically designed to detect defects in an array containing repeating structural elements. 【0065】 As shown in Figure 2A, one or more arrays 260 are separated from one or more regions 265 by one or more boundaries 266. The boundaries 266 define the physical limits between the arrays 260 and the corresponding surrounding regions 265. 【0066】 According to some embodiments, this method enables the estimation of one or more first regions 270 of an image 250 that include only one or more arrays 260 up to the boundary 266. In particular, according to some embodiments, this method enables the identification of arrays 260 up to the boundary 266, except for one or more second regions 275 corresponding to one or more regions 265. 【0067】 Let us focus on Figure 3, which shows a non-limiting embodiment of operations 210 to 230. 【0068】 As described with reference to operation 210, a correlation matrix is ​​obtained. A non-restrictive example of a correlation matrix 365 obtained in a given region of image 250 is shown in Figure 3A. 【0069】 Therefore, this method may include determining a subregion of image 250 corresponding to a value in the correlation matrix 365 that satisfies an amplitude criterion (operation 300). The amplitude criterion may define, for example, a subregion of image 250 where a local maximum correlation peak is identified (in some embodiments, an absolute threshold may be set) as corresponding to the location of a repeating structural element in image 250. According to some embodiments, during the setup phase prior to the runtime test of the sample, a first estimate of the amplitude of the correlation peak obtained in the subregion containing one of the repeating structural elements can be used to determine the amplitude criterion to be used during the runtime test, the amplitude criterion taking into account the presence of the structural element. 【0070】 As shown in Figure 3A, the correlation matrix 365 contains correlation peaks (maxima) 367 located in subregions 375 of the image. These subregions 375 correspond to the estimation of the locations of repeating structural elements. In fact, since the correlation analysis involves correlating the pixel intensity of the image with the pixel intensity of a reference image that provides information about the repeating structural elements, it is expected that subregions of the image containing repeating structural elements will provide higher correlation values ​​compared to subregions (region 265) of the image that do not contain repeating structural elements. 【0071】 This method may further include clustering the sub-region 375 into one or more clusters (operation 310) based on data that provides information about the distances between repeating structural elements in the array. 【0072】 As described above, repeating structural elements are generally arranged according to a repeating pattern or grid. Therefore, it is possible to obtain the distance between two consecutive structural elements in the array. In some embodiments, data can be obtained that gives information about the distance between repeating structural elements 361 in the array along a first axis (e.g., the horizontal axis, corresponding to the rows of an image) (see reference 368 in Figure 3B), and data that gives information about the distance between repeating structural elements in the array along a second axis (e.g., the vertical axis, corresponding to the columns of an image) (see reference 369 in Figure 3B). As shown in the illustration, the distance can be measured between the centers of the structural elements. 【0073】 According to some embodiments, operation 310 may include clustering subregions 375 into one or more first clusters based on data that provides information about the distances between repeating structural elements in an array along a first axis 372. According to some embodiments, in a given cluster, subregions 375 are located at a distance from another subregion 375 in the cluster that is less than or equal to the distance between repeating structural elements along the first axis. A non-limiting example in which subregions are assigned to the same cluster 370 along the first axis 372 is shown in Figure 3A. As shown, subregion 374 is not assigned to cluster 370 because the distance from each subregion to cluster 370 exceeds the distance between two repeating structural elements along the first axis 372. 【0074】 According to some embodiments, operation 310 may include clustering subregions into one or more second clusters based on data that provides information about the distances between repeating structural elements in an array along a second axis 373. According to some embodiments, in a given cluster, a subregion is located at a distance from another subregion of the cluster that is less than or equal to the distance between repeating structural elements along the second axis 373. A non-limiting example in which subregions are assigned to the same cluster 381 along the second axis 373 is shown in Figure 3A. As shown, subregion 383 is not assigned to cluster 381 because the distance from subregion 383 to cluster 381 exceeds the distance between two repeating structural elements along the second axis 373. 【0075】 This method includes determining one or more first regions based on at least one or more clusters (operation 320). In particular, a first cluster can be used to determine the size and location of one or more first regions along a first axis 372, and a second cluster can be used to determine the size and location of one or more first regions along a second axis 373. For example, cluster 370 provides the size and location of a first region along axis 372, and cluster 381 intersecting cluster 370 provides the size and location of the same first region along axis 373. As a result, a first region 384 is identified. This can be done for all clusters, and all clusters can be used to determine the boundaries of different first regions. 【0076】 According to some embodiments, another operation is performed to identify a first area using clusters. In particular, this method may include determining a polygon (e.g., a rectangle or a square) enclosing one or more clusters identified to define a first area, and outputting the polygon as the first area. For example, in the example of Figure 3A, a rectangle 392 can be generated that encloses the first area identified based on a first cluster 370 and a second cluster 381. 【0077】 Next, we consider Figure 3C. According to some embodiments, this method can include selecting only clusters in which some sub-regions present satisfy a threshold (e.g., exceed the threshold) (operation 330). This is shown in Figure 3C. Multiple clusters 3831 to 3835 are identified along axis 372. The cluster referred to as 3831 contains only one sub-region 384. Since we know that the array contains repeating structural elements arranged along a repeating pattern (e.g., a grid), we can assume that sub-region 384 does not correspond to a structural element because the repeating pattern does not contain any isolated structural elements. Therefore, this cluster can be ignored or removed when determining one or more first regions in operation 320. The same can be applied to clusters determined along a second axis 373 (second clusters). If a given cluster contains some sub-regions that are below the threshold, the given cluster is ignored when determining one or more first regions in operation 320. 【0078】 Next, let us look at Figure 4. According to some embodiments, a reference image that provides information on at least one of the repeating structural elements can be processed using an image processing algorithm. According to some embodiments, the image processing algorithm attenuates the repeating pattern of the reference image. For example, partial whitening can be applied to the reference image. Partial whitening involves, for example, converting the reference image to the frequency domain (e.g., converting X(i,j) representing pixels in the reference image to X(f) in the frequency domain) and degrading high / strong frequencies (e.g., 【0079】 【number】 This may include performing an inverse transform to return to the image (for example, transforming X'(f) to X'(i,j)). 【0080】 A non-limiting example of the method in Figure 4 is shown in Figure 4A. As shown, the array 460 includes repeating structural elements 410 and conductive lines 411. The conductive lines 411 extend to a region 465 surrounding the array 460. A reference image 470 is obtained that provides information about the repeating structural elements 410, as shown in Figure 4B. To obtain a modified reference image 475, image processing (e.g., partial whitening) that attenuates the repeating pattern is applied to the reference image 470. As shown, both the conductive lines 411 (corresponding to the repeating pattern) and the structural elements 410 (also corresponding to the repeating pattern) are attenuated in the modified reference image 475. As a result, when correlation is performed between the modified reference image 475 and the image (as described with reference to operation 210), both the sub-regions and regions, in this embodiment, contain common repeating features (conductive lines 411), but the sub-regions corresponding to the structural elements will provide a higher correlation value than the sub-regions corresponding to the regions, thereby making it easier to distinguish the array from the surrounding region. 【0081】 According to some embodiments, the method may include acquiring an image of a sample including one or more arrays and one or more surrounding regions (operation 500). Operation 500 is similar to operation 200 described above. In some embodiments, the image is acquired by an electron beam test tool. In some cases, the signal-to-noise ratio of the image may be low, and therefore the method shown in Figure 2, which requires correlation with a reference image, may not always be applicable. A low signal-to-noise ratio may be due to the size of feature areas present in the sample, charging effects, etc. In some embodiments, due to a low signal-to-noise ratio, structural elements of the array cannot be identified / distinguished within the array by visual inspection of the image. In some embodiments, the size of the pixels in the image may be larger than the size of the structural elements, and therefore, the structural elements cannot be distinguished by visual inspection. 【0082】 This method provides data D that gives information about the pixel intensity of at least one of the array and the surrounding region. threshold This further includes obtaining (operation 510). thresholdThis can be acquired before the runtime test of the sample, particularly during the setup phase. For example, during the setup phase, an image of the sample similar to the sample under runtime test is acquired. An operator or automated algorithm (e.g., K-means algorithm) provides an initial estimate of the location of the array and surrounding region in the image. The mean value P of the pixel intensity of the array. array The average value P of the pixel intensity in the surrounding area is calculated. region These are calculated. These two values ​​are expected to be different because the array and the surrounding region contain different structural features. D threshold For example, P array and P region It can be calculated based on D. threshold For example, this can correspond to the average between these two values, but this is not limited. 【0083】 This method further includes determining data (operation 520) that represent the pixel intensity along multiple axes of the image. In particular, this includes data D that represents the pixel intensity along each of multiple rows of the image. X And data D, which represents the pixel intensity along each of the multiple columns of the image. Y This may include determining the following. Data D X (or D Y This can be calculated, for example, as the average value of the pixel intensity along each row (or column) of the image. 【0084】 This method distinguishes between one or more first regions of an image corresponding to one or more arrays and one or more second regions of an image corresponding to one or more regions, D X , D Y , and D threshold This further includes using (operation 530). 【0085】 Operation 530 is D X Threshold D threshold Rows of an image that are greater than (or less than, depending on whether the pixel intensity is higher with respect to the array or surrounding area) (for example, acquired during the setup phase), and DY Threshold D threshold This may include identifying columns of images that exceed (for example, those acquired during the setup phase). The intersection of the identified rows and columns will then be used to identify the location of the array. 【0086】 A non-limiting example is provided in Figure 5A. For example, during the setup phase, the pixel intensity of the array is (on average) higher than that of the surrounding area (P array P region It has been determined that (it is greater than) D threshold P array and P region Assume that it is set as the average value. Data D gives information about the pixel intensity along the row. X This is shown as curve 545 (this curve is purely illustrative and not limiting). As shown in the figure, in the row of the image where array 562 is located, the curve is threshold D threshold It exceeds (referenced as 548). Data D gives information about pixel intensity along the column. Y This is shown as curve 561 (this curve is purely illustrative and not limiting). As shown in the figure, in the row of images in which array 562 is placed, the curve is at threshold D threshold It exceeds (referenced as 548). 【0087】 This method further includes outputting data (operation 540) that provides information about one or more first regions of an image. Operation 540 is similar to operation 230 described above. In particular, according to some embodiments, the location of the first region corresponding to the location of the array is estimated by the intersection of a row where curve 545 exceeds a threshold 548 and a column where curve 561 exceeds a threshold 548. 【0088】 According to some embodiments, at least operations 510, 530, and 540 are performed during the runtime scanning of the sample. In other words, the method for identifying the array in the image is efficient and therefore can be performed during the runtime phase. 【0089】 Next, consider Figure 6. In some cases, the data representing pixel intensity along the row and / or column of the image in which the array is located may be close to the data representing pixel intensity along other rows and / or columns. A non-limiting example in which array 660 is located in the lower-left corner of the image and surrounded by a large area 665 is shown in Figure 6A. The method in Figure 6 is a possible embodiment of a solution that can improve the distinction between the row and column of the image in which the array is located and other rows and columns. 【0090】 This method uses data D, which represents the pixel intensity along each of multiple rows of the image (for example, the average of the pixel intensity along the row). X This includes determining (operation 610). X However, this is represented as curve 668 in Figure 6A. The first threshold D threshold,1 Assume that it has been obtained (for example, during the setup phase before runtime testing). threshold,1 This provides information about the pixel intensity of at least one of the array and region. In some embodiments, D threshold,1 The mean value of the pixel intensity of the array can be selected as a strict threshold (high threshold) to maximize the probability of distinguishing between rows of the image corresponding to the array and other rows. For example, during the setup phase (performed on images of samples similar to the sample under test during runtime), the mean value of the pixel intensity of the array P array The average value P of the pixel intensity in the surrounding area is calculated. region Let's assume that has been calculated (as explained above). For example, P array P region Assume it is higher than D threshold,1 P region By selecting a value higher than this, the probability of removing rows corresponding to the surrounding region can be maximized. For example, D threshold,1 The following can be selected: D threshold,1 =P region +N*σ (where σ is the standard deviation of the pixel intensity in the surrounding region, and N is an integer equal to, for example, 2). 【0091】 This method is D X is D threshold,1A subset of images S that includes rows of images exceeding a certain limit. L This includes selecting (represented as 682). Subset S, as shown in the figure. L This includes row 683 of the image where array 660 is located, and an additional row 684 of the image that does not contain array 660 (however, the pixel intensity of these additional rows is D threshold,1 This method includes (which exceeds) and. This method is subset S L Data D represents the pixel intensity along each of the multiple columns (for example, the average of the pixel intensity along the column). Y,SL This further includes determining (operation 630) the curve 686 in Figure 6B. Y,SL S exceeds the second threshold of 690. L Subset C of the column SL This includes determining (operation 640) (referenced as 689). This second threshold 690 can be obtained based on measurements taken during the setup phase before the runtime test. For example, the second threshold is P array (average pixel intensity of the array) and P region It can be set to be equal to the average value of (the average pixel intensity of the surrounding area). However, this is not limited. 【0092】 Column C SL (Reference 689) indicates the position of the array along the row axis. The position and size of the array along the column axis (Y axis) of the image can then be determined, thereby yielding one or more first regions in the image corresponding to the array (operation 650). The positions of one or more first regions in the image (corresponding to the array) can be provided, and / or the positions of one or more second regions in the image (corresponding to the surrounding region) (corresponding to all regions not identified as first regions). 【0093】 In fact, once the row 689 of the image corresponding to the array is identified, it becomes easier to distinguish between the row of the image containing the array and the other rows of the image, as is clearly visible in Figure 6C. (Image subset S') L(referred to as 692 in FIG. 6C) is conceivable. This subset S’ L includes all rows of the image and is limited to the columns 689 of the image identified in the previous operation. This method determines data (referred to as 693) representing pixel intensities (e.g., average pixel intensity) along each of the multiple rows of the subset S’ L (692), and further includes determining rows 694 of the image in which the data 693 exceeds a third threshold 695 (in some embodiments, the third threshold 695 is equal to the second threshold 690, but this is not essential). As can be clearly seen in FIG. 6C, now it becomes easier to distinguish the rows of the image including the array from other rows based on pixel intensity. These rows 694, together with the columns 689, define one or more first regions of the image corresponding to the array. The other regions of the image correspond to the second regions of the image corresponding to the area surrounding the array. 【0094】 In the example from FIG. 6 to FIG. 6C, this method starts by selecting a subset S of the image that includes rows whose average pixel intensity exceeds a threshold. L It should be understood that this method can be executed in the same way by first selecting a subset of columns. In this case, this method - determines data D representing pixel intensities along each of the multiple columns of the image (corresponding to operation 610), Y - selects a subset S of the image that includes columns of the image in which D exceeds a first threshold (corresponding to operation 620), Y - determines data D representing pixel intensities along each of the multiple rows of the subset S C (corresponding to operation 630), - determines a subset L of the rows of S C in which D X,SC exceeds a second threshold (corresponding to operation 640), - D X,SC where D exceeds a second threshold (corresponding to operation 640), C and determines a subset L of the rows of S SC (corresponding to operation 640), - L SCThis may include determining one or more first regions corresponding to the array based on at least the following (corresponding to operation 650 - since the rows of the image corresponding to the array are known, it becomes easier to identify the columns of the image corresponding to the array, as described with reference to Figure 6C). 【0095】 According to some embodiments, this method further enables the estimation of one or more first regions of an image containing only one or more arrays up to a boundary separating the arrays from the surrounding region. In particular, according to some embodiments, this method enables the identification of arrays up to the boundary, except for one or more second regions corresponding to one or more regions. 【0096】 Next, let us consider Figure 7. Assume that an image of the sample has been acquired and one or more first regions have been identified, each corresponding to one or more arrays 710, each containing repeating structural elements 720. This identification can rely, for example, on the various embodiments described above, or on other identification methods. Therefore, an image 700 limited to one or more first regions (corresponding to arrays) is available without the region surrounding one or more arrays. In some embodiments, the image 700 includes both one or more first regions (corresponding to arrays) and one or more second regions (corresponding to regions). Since the location of the first regions is known, it is possible to manipulate only the first regions. Hereafter, we will refer to an image 700 containing only the first regions (corresponding to arrays), but it should be understood that this method can be similarly applied to images containing both the first and second regions by applying this method only to the first regions of the image. 【0097】 As is clearly visible in Figure 7, in some embodiments, the image 700 of the array is distorted. In particular, the positions of the structural elements 720 in the array, which are clearly visible in image 700, do not match the intended positions in the array (their true positions in the sample). This may be due to various factors, such as measurement errors in the test tools. 【0098】 Distortion can be a problem when attempting to use image 700 for various applications such as defect detection and / or classification. Therefore, it is necessary to correct this distortion. Figure 8 shows one embodiment of a method for correcting distortion present in the array image. 【0099】 This method includes performing a correlation analysis (operation 800) between the pixel intensity of image 700 and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements. The reference image used in operation 800 may be different from the reference image used in operation 210 to identify a first region of the image corresponding to the array (in this case, a second reference image different from the first reference image used in operation 210 is used in operation 800). However, this is not required. The output of the correlation analysis is a second correlation matrix (which may be different from the correlation matrix obtained in operation 210). In some embodiments, it is possible to reuse the correlation matrix obtained in operation 210 (in this case, only the values ​​corresponding to one or more first regions are used). 【0100】 This method may further include determining subregions of an image corresponding to values ​​of a second correlation matrix that satisfy an amplitude criterion (operation 810). In particular, the amplitude criterion may require that subregions of an image associated with maximal values ​​(e.g., local maximal values) of the second correlation matrix be identified. 【0101】 As shown in Figure 8A, the second correlation matrix 860 contains correlation peaks (maxima) located in a given subregion 685. These subregions 865 correspond to the estimation of the location of the repeating structural elements (in particular, the central region of each structural element). In fact, since the correlation analysis involves correlating the pixel intensity of the image with the pixel intensity of a reference image that provides information about the repeating structural elements, it is expected that subregions of image 700 containing repeating structural elements will provide higher correlation values ​​compared to subregions of image 700 that do not contain repeating structural elements. 【0102】 This method may further include determining a distortion map between image 700 and the array (operation 820). The distortion map can be determined based on data that gives information about the locations of subregions (determined using a second correlation matrix) and the intended locations of repeating structural elements in the array. 【0103】 Figure 8B shows a non-restrictive example indicating the location of sub-region 865 corresponding to the maxima of the second correlation matrix, and the planned location 866 of the structural element in the array. For each sub-region, it is possible to determine a distortion vector 867 that shows the difference between the location of the structural element in the image (estimated using the maxima of the second correlation matrix) and the planned location 866 of the corresponding structural element. 【0104】 According to some embodiments, the distortion map can be determined for the entire image. In fact, as described above, the distortion between the location of the subregion 865 in the image and the data giving information about the intended location of the repeating structural elements in the array (see 867, hereafter referred to as "DF") is determined. central The ) is determined. This corresponds to the distortion of the central part of each structural element relative to the intended position. To determine the distortion of other pixels in the image (which do not necessarily correspond to the central part of the structural element), this method is used across the image. centralThis may include applying an interpolation method to the values. This allows for an estimation of the distortion of all other pixels located between different sub-regions 865. According to some embodiments, the interpolation method is applied separately to the distortion along the X-axis (rows of the image) and the distortion along the Y-axis (columns of the image). 【0105】 This method may further include generating a corrected image 880 (see Figure 8C) based on a distortion map (operation 830). This may include moving pixels in the image based on the distortion map so that the positions of subregions 865 in the corrected image 880 (corresponding to peaks in the second correlation matrix) and the positions of data giving information about the intended positions of repeating structural elements in the array satisfy a proximity criterion (e.g., the difference in positions is less than a threshold). 【0106】 Next, let us look at Figure 9. According to some embodiments, the method may include obtaining a reference image that provides information on at least one repeating structural element (900) and selecting only a subset of the reference image as a second reference image (910). According to some embodiments, the size of the subset is based on a compromise. On the one hand, the size of the subset must be large enough to identify the positions of the structural elements in the image, and on the other hand, the size of the subset must be small enough to obtain a sufficient number of correlation values. 【0107】 A non-restrictive example is shown in Figure 9A. 【0108】 Reference image 920 has been acquired. A subset 930 of reference image 920 is selected. This subset can be used as a second reference image in the method shown in Figure 8. 【0109】 According to some embodiments, the subset 930 can be selected using an iterative method, for example, during the setup phase. This method begins with a first subset (the maximum size of this subset may be set, for example, by the user). The method in Figure 8 is performed using this first subset. The resolution is then improved, which means the size of the first subset is reduced. The method in Figure 8 is performed again using this new subset, and the performance of the output is compared to the previous iteration. If the performance is improved, the method is repeated with a new subset of a smaller size. If the performance is not improved, the method is stopped, and the subset obtained in the previous iteration is selected. 【0110】 It should be understood that the present invention is not limited in its application to the details described herein or shown in the drawings. 【0111】 It will also be understood that the system according to the present invention can be implemented, at least in part, on a properly programmed computer. Similarly, the present invention envisions a computer program readable by a computer to perform the method of the present invention. The present invention further envisions a non-temporary computer-readable memory that explicitly embodies a program of instructions executable by a computer to perform the method of the present invention. 【0112】 Other embodiments of the present invention are possible and can be practiced and implemented in various ways. Therefore, it should be understood that the terminology and language used herein are for illustrative purposes only and should not be considered limiting. Accordingly, those skilled in the art will understand that the concepts on which this disclosure is based can be readily used as a basis for designing other structures, methods, and systems to accomplish some of the objectives of the subject matter of this disclosure. 【0113】 Those skilled in the art will readily understand that various modifications and changes can be applied to the embodiments of the invention described above without departing from the scope of the appended claims and the scope defined by the appended claims. [Explanation of symbols] 【0114】 100 Test Systems 101 Low-Resolution Testing Tools 102 High-Resolution Testing Tools 103 Computer-based systems, operating systems 104 Processor and Memory Circuit (PMC) 105 Hardware-based input interfaces 106 Hardware-based output interfaces 107 Storage Systems 108. Graphical User Interface (GUI) 109 Data repositories 110 CAD Server 112 Deep Neural Networks (DNNs) 121 Low-resolution image data 122 High-resolution image data 123 Instruction-related data 124 Instruction-related data 250 images 260 repeating structural elements, arrays 265 areas 266 Boundary 270 First area 275 Second area 361 Repetitive structural elements 365 Correlation Matrix 367. Correlation peak (maximum value) See 368. See 369. 370 clusters 372 The first axis 373 The second axis 374 Sub-areas 375 Sub-areas 381 clusters Clusters 3831-3835 383 Sub-areas 384 Sub-areas 392 Rectangle 410 Repetitive structural elements 411 Conductive Line 460 arrays 465 area 470 Reference Image 475 Modified reference image 545 Curve 548 threshold 561 Curve 562 array 660 array 665 Large area 668 curve 683 rows of images 684 Additional row of images 685 Sub-areas 686 curve 689 columns 690 Second threshold A subset of 692 images 693 data 694 lines 695 Third threshold 700 images 710 Array 720 Repeating structural elements, structural elements 860 Second correlation matrix 865 Sub-area 866 Planned location 867 Distortion vector 880 Modified image 920 Reference Image 930 subset

Claims

[Claim 1] Each of the arrays contains a repeating structural element, Each region comprises one or more regions that at least partially enclose the corresponding array and include a feature distinct from the repeating structural element, A system comprising a processor and memory circuit (PMC) configured to acquire an image of a semiconductor sample including, The PMC, during the runtime scanning of the semiconductor sample, To obtain a correlation matrix, a correlation analysis is performed between the pixel intensity of the image and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements. The correlation matrix is ​​used to determine the sub-region of the aforementioned image, Based on data that provides information on the distance between the repeating structural elements in the array, the sub-regions are clustered into one or more clusters. The one or more clusters are used to distinguish between one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more areas. The system outputs data that provides information about one or more first regions of the aforementioned image. A system configured in such a way. [Claim 2] The aforementioned PMC is Based on data that provides information on the distance between the repeating structural elements in the array along the first axis, the sub-regions are clustered into one or more first clusters. Based on data that provides information on the distance between the repeating structural elements in the array along the second axis, the sub-regions are clustered into one or more second clusters. The first cluster and the second cluster are used to distinguish between one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more areas. It is further configured in the following way: The system according to claim 1. [Claim 3] The sub-regions of the aforementioned image correspond to the values ​​of the correlation matrix that satisfy the amplitude criterion, The system according to claim 1. [Claim 4] The one or more arrays are separated from the one or more regions by one or more boundaries, The PMC is further configured to estimate one or more first regions of the image, which include only the at least one or more arrays up to the boundary. The system according to claim 1. [Claim 5] The PMC is configured to apply image processing to the reference image, The aforementioned image processing attenuates the repeating pattern of the reference image. The system according to claim 1. [Claim 6] The aforementioned PMC, for each cluster, Define a polygon that encloses one or more of the aforementioned clusters, The polygon is further configured to be output as the first region of the image. The system according to claim 1. [Claim 7] The one or more clusters are configured to include only clusters where some sub-regions satisfy the threshold. The system according to claim 1. [Claim 8] The aforementioned PMC is To obtain a second correlation matrix, a correlation analysis is performed between the pixel intensity of one or more first regions of the image and the pixel intensity of a second reference image that provides information on at least one of the repeating structural elements. Determine a given sub-region in one or more first regions of the image that corresponds to the value of the second correlation matrix that satisfies the amplitude criterion. Based at least on the location of the given sub-region within the one or more first regions of the image and data providing information on the planned location of the repeating structural elements in the array, a distortion map between the one or more first regions of the image and the array is determined. Based on the aforementioned distortion map, it is further configured to generate a corrected image. The system according to claim 1. [Claim 9] The PMC is further configured to acquire a reference image that provides information on at least one of the repeating structural elements, and to select only a subset of the reference image as the second reference image. The system according to claim 8. [Claim 10] The aforementioned PMC is The distortion DFcentral is determined between the position of a given sub-region in one or more first regions of the image and data providing information about the planned position of the repeating structural elements in the array. The system is further configured to determine the map of the distortion between the one or more first regions of the image and the array of the semiconductor sample, based on an interpolation method applied to at least DFcentral. The system according to claim 8. [Claim 11] The aforementioned PMC is The data providing information on the location of the given sub-region in the modified image and the planned location of the repeating structural element in the array is further configured to generate the modified image such that a proximity criterion is met. The system according to claim 8. [Claim 12] The aforementioned PMC is The system is further configured to acquire data that provides the amplitude reference information during the setup phase prior to the runtime test of the semiconductor sample. The system according to claim 8. [Claim 13] The aforementioned PMC is The data providing information on the location of the sub-regions in the modified image and the planned location of the repeating structural elements in the array is further configured to generate the modified image such that a proximity criterion is met. The system according to claim 8. [Claim 14] A non-temporary computer-readable medium containing a program of instructions, which, when executed by one or more computers, will be used by the one or more computers. Each of the arrays contains a repeating structural element, Each region comprises one or more regions that at least partially enclose the corresponding array and include a feature distinct from the repeating structural element, To obtain an image of a semiconductor sample containing, To obtain a correlation matrix, a correlation analysis is performed between the pixel intensity of the image and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements. The correlation matrix is ​​used to define subregions of the aforementioned image, The sub-regions are clustered into one or more clusters based on data that provides information on the distances between the repeating structural elements in the array, The one or more clusters are used to distinguish between one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more areas. Outputting data that provides information about one or more first areas of the aforementioned image, A non-temporary computer-readable medium that enables execution. [Claim 15] The sub-regions of the aforementioned image correspond to the values ​​of the correlation matrix that satisfy the amplitude criterion, The non-temporary computer-readable medium according to claim 14. [Claim 16] When executed by the one or more computers, the one or more computers will The sub-regions are clustered into one or more first clusters based on data that provides information on the distance between the repeating structural elements in the array along the first axis, The sub-regions are clustered into one or more second clusters based on data that provides information on the distance between the repeating structural elements in the array along the second axis, The first cluster and the second cluster are used to distinguish between one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more areas. A non-temporary computer-readable medium according to claim 14, which causes the execution of the following. [Claim 17] When executed by the one or more computers, the one or more computers will To obtain a second correlation matrix, a correlation analysis is performed between the pixel intensity of one or more first regions of the image and the pixel intensity of a second reference image that provides information on at least one of the repeating structural elements. Determine a given sub-region in one or more first regions of the image that corresponds to the value of the second correlation matrix that satisfies the amplitude criterion. Based at least on the location of the given sub-region within the one or more first regions of the image and data providing information on the planned location of the repeating structural elements in the array, a distortion map between the one or more first regions of the image and the array is determined. Based on the aforementioned distortion map, a corrected image is generated. A non-temporary computer-readable medium according to claim 14, which causes the execution of the following. [Claim 18] The sub-regions of the aforementioned image correspond to the values ​​of the correlation matrix that satisfy the amplitude criterion, The non-temporary computer-readable medium according to claim 14. [Claim 19] The one or more arrays are separated from the one or more regions by one or more boundaries, A non-temporary computer-readable medium containing a program of instructions, which, when executed by one or more computers, causes the one or more computers to estimate one or more first regions of the image, which include only the at least one or more arrays up to the boundary. The non-temporary computer-readable medium according to claim 14. [Claim 20] A method that can be implemented using a computer. Each of the arrays contains a repeating structural element, Each region comprises one or more regions that at least partially enclose the corresponding array and include a feature distinct from the repeating structural element, To obtain an image of a semiconductor sample containing, To obtain a correlation matrix, a correlation analysis is performed between the pixel intensity of the image and the pixel intensity of a reference image that provides information on at least one of the repeating structural elements. The correlation matrix is ​​used to define subregions of the aforementioned image, The sub-regions are clustered into one or more clusters based on data that provides information on the distances between the repeating structural elements in the array, The one or more clusters are used to distinguish between one or more first regions of the image corresponding to the one or more arrays and one or more second regions of the image corresponding to the one or more areas. Outputting data that provides information about one or more first areas of the aforementioned image, A method that includes [a certain feature].

Images