Atomic array fluorescence image automatic identification method and device

Abstract

This invention provides an automated method and apparatus for identifying atomic array fluorescence images. By averaging the pixels of multiple atomic fluorescence images to obtain an average fluorescence image, and segmenting regions based on differences in pixel values ​​within the average fluorescence image, the location and size of the fluorescence regions where atoms may be located are determined. This allows for the precise acquisition of the locations where captured atoms may exist on the average fluorescence image. The presence of an atom at that location is determined by the fluorescence intensity of each atomic fluorescence image at that location. Through the superposition and calculation of multiple atomic fluorescence images, the unreliability of captured atoms under low loading rates is overcome, and the influence of noise interference is avoided, improving the accuracy and reliability of atom identification in the corresponding scenarios.

Description

Technical Field

[0001] This invention belongs to the field of quantum computing, and more specifically, relates to an automated method and apparatus for recognizing fluorescent images of atomic arrays. Background Technology

[0002] In neutral atom quantum computing, optical trap arrays are commonly used to achieve large-scale trapping and manipulation of atoms. Optical trap arrays can be generated using photoelectric modulation devices (such as spatial light modulators), and their holograms can be calculated using corresponding algorithms. The array configuration can be customized according to requirements. However, when loading atoms into magneto-optical traps with laser cooling, the number of atoms loaded into each optical trap is either 0 or 1 due to collision blocking effects. The loading process is random, with a typical loading rate below 0.5, making it impossible to guarantee successful loading of atoms every time. Atomic clusters are loaded into corresponding optical trap arrays, collected and amplified, and then detected by a camera. This yields images of randomly loaded atoms within the optical trap array, allowing for the identification of the atom arrangement and loading state by observing the fluorescence distribution on the images. For experiments involving atom rearrangement, the precise distribution of atoms is crucial, necessitating automated identification and localization of the atomic fluorescence images. However, current technology suffers from the challenge that the fluorescence emitted by a single atom covers multiple pixels in an image and exhibits irregular shapes, making it difficult to accurately pinpoint its location and intensity using methods like Gaussian fitting. This results in insufficient accuracy in atom identification on the atomic fluorescence images, impacting the accuracy of subsequent computational applications.

[0003] Therefore, overcoming the shortcomings of the existing technology is an urgent problem to be solved in this technical field. Summary of the Invention

[0004] The problem this invention aims to solve is how to improve the accuracy of atomic identification in atomic fluorescence images.

[0005] In a first aspect, an automated recognition method for atomic array fluorescence images is provided, comprising: Multiple atomic fluorescence images were acquired during atomic loading, and pixel averaging was performed to obtain an average fluorescence image. The average fluorescence image is denoised to obtain a denoised image. The region of fluorescent pixels in the denoised image is segmented to obtain multiple fluorescent regions, and the average region radius of the fluorescent regions is obtained. Based on the type of optical trap, obtain the theoretical atomic positions on the denoised image; Based on the average region radius and the theoretical atom position, a fluorescence intensity dataset of all atomic fluorescence images at the corresponding theoretical atom positions is obtained. Based on the fluorescence intensity dataset, a representative value of the intensity at the corresponding theoretical atom position is set, and the presence of an atom at the theoretical atom position is determined based on the average of the representative values ​​of the intensity at all theoretical atom positions.

[0006] Preferably, the step of denoising the average fluorescence image to obtain a denoised image specifically includes: The average fluorescence image is used as the input image in the initial iteration round, and each iteration round includes: Calculate the pixel value of all pixels in the input image. Arithmetic mean ; Generate an optimized image, optimizing the pixel value of each pixel in the image. Pixels with a value greater than 0 are identified as fluorescent pixels, and the number of fluorescent pixels in the optimized image is obtained. The optimized image is linearly normalized to a preset interval to obtain the output image; Determine whether the number of fluorescent pixels in this iteration satisfies the convergence condition; If the convergence condition is met, the output image is used as the denoised image; If the convergence condition is not met, the output image is used as the input image in the next iteration.

[0007] Preferably, the step of segmenting the region of fluorescent pixels in the denoised image to obtain multiple fluorescent regions and obtaining the average region radius of the fluorescent regions specifically includes: Obtain the set of locations of all fluorescent pixels in the denoised image; Treat all adjacent fluorescent pixels as a single fluorescent region; Obtain the region radius of each fluorescent region in the denoised image, and then obtain the average region radius of the denoised image.

[0008] Preferably, when the atomic assembly rate of the captured optical trap is higher than or equal to a preset intensity representative value, the step of obtaining the theoretical atomic positions on the denoised image according to the type of captured optical trap specifically includes: The intensity centroid coordinates of each fluorescent region are calculated using the following expression: ; ; in, The x-coordinate of the intensity centroid coordinates The ordinate of the intensity centroid coordinates Let (x, y) be the set of all pixels contained in the i-th region, and (x, y) be the set of positions of these pixels. The image light intensity value of the pixel located at (x,y).

[0009] Preferably, when the atomic assembly rate of the captured optical trap is lower than a preset intensity representative value, the step of obtaining the theoretical atomic positions on the denoised image according to the type of captured optical trap specifically includes: Obtain the optimal rotation angle of the denoised image, convert the denoised image into an aligned image based on the optimal rotation angle, and obtain the pixel values ​​of all fluorescent pixels in the aligned image; Obtain the row mean of pixel values ​​for each row in the aligned image. Perform local peak detection on the row mean of all rows in the aligned image and select the row where the row mean is located at the local peak position as the candidate row center. When there are other candidate row centers within the interval of the positive and negative average regions in the vertical direction, retain the candidate row center with the highest row mean and remove the other candidate row centers. Obtain the column mean of pixel values ​​in each column of the aligned image. Perform local peak detection on the column mean of all columns in the aligned image. Columns whose column mean is located at the local peak position are selected as candidate column centers. When there are other candidate column centers within the interval of the positive and negative average regions in the horizontal direction, retain the candidate column center with the highest column mean and remove the other candidate column centers. By combining all candidate row centers and candidate column centers, the rotational atom positions in the aligned image are obtained. All rotational atom positions are then subjected to inverse coordinate transformation according to the optimal rotation angle to obtain the theoretical atom positions of all atoms in the denoised image.

[0010] Preferably, obtaining the optimal rotation angle for the denoised image specifically includes: Within a preset angle range, the denoised image is rotated multiple times around the center position of the image according to the first preset angle step size, and a rotated image is obtained after each rotation; The coordinates of each fluorescent pixel in the rotated image in the denoised image are obtained as adjustment coordinates. The pixel values ​​of the four adjacent related pixels at the adjustment coordinates are obtained. The pixel value of the corresponding fluorescent pixel in the rotated image is calculated based on the pixel values ​​of the related pixels. Based on the pixel values ​​of all fluorescent pixels in the rotated image, obtain the row pixel mean of each row of pixels in the rotated image, obtain the column pixel mean of each column of pixels in the rotated image, and obtain the variance sum of the rotated image based on the row pixel mean of each row of pixels and the column pixel mean of each column of pixels. The rotation angle with the largest variance within a preset angle range is selected as the preferred angle. With the preferred angle as the center, multiple rotations are performed within a preset positive and negative interval with a second preset angle step size. The variance of the rotation image for each rotation is obtained, and the rotation angle with the largest variance is selected as the optimal rotation angle.

[0011] Preferably, the step of calculating the pixel value of the corresponding fluorescent pixel in the rotated image based on the pixel value of the associated pixel specifically includes: The expression for calculating the pixel value of the corresponding fluorescent pixel in the rotated image is: ; in, , , To adjust the x-coordinate of the coordinate system, To adjust the ordinate of the coordinate system, The associated pixel is located in the bottom left corner. The associated pixel is located in the bottom right corner. The associated pixel is located in the top left corner. The associated pixel is located in the upper right corner.

[0012] Preferably, obtaining the fluorescence intensity dataset of all atomic fluorescence images at the corresponding theoretical atomic positions based on the average region radius and the theoretical atomic positions specifically includes: For each atomic fluorescence image, a square region with a side length of twice the average region radius is defined as the calculation region, centered on the theoretical atom position; Obtain all intersection pixels that intersect with the corresponding calculation region, and use the overlap area between the intersection pixels and the calculation region as the contribution weight of the corresponding intersection pixels; The pixel values ​​of all intersecting pixels are weighted and summed according to their respective contribution weights to obtain the fluorescence feature intensity of the corresponding theoretical atomic position in the atomic fluorescence image. The fluorescence characteristic intensities of all atomic fluorescence images at the corresponding theoretical atomic positions are obtained and used as a fluorescence intensity dataset at the corresponding theoretical atomic positions.

[0013] In a second aspect, an automated recognition device for atomic array fluorescence images is provided, comprising at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the processor to perform the automated recognition method for atomic array fluorescence images.

[0014] Thirdly, the present invention also provides a non-volatile computer storage medium storing computer-executable instructions that are executed by one or more processors to perform the method described in the first aspect.

[0015] Fourthly, a chip is provided, comprising: a processor and an interface for calling and running a computer program stored in memory, performing the method as described in the first aspect.

[0016] Fifthly, a computer program product containing instructions is provided that, when executed on a computer or processor, causes the computer or processor to perform the method as described in the first aspect.

[0017] In a sixth aspect, an automated recognition system for atomic array fluorescence images is provided, comprising an automated recognition device for atomic array fluorescence images as described in the second aspect, and using an automated recognition method for atomic array fluorescence images as described in the first aspect.

[0018] Unlike existing technologies, the present invention has at least the following beneficial effects: By averaging the pixels of multiple atomic fluorescence images, an average fluorescence image is obtained. Based on the differences in pixel values ​​within the average fluorescence image, region segmentation is performed to determine the location and size of the fluorescence region where the atom might be located. This allows for the precise determination of the possible locations of trapped atoms on the average fluorescence image. The presence of an atom at that location is then determined by the fluorescence intensity of each atomic fluorescence image at that location. This method of superimposing and calculating multiple atomic fluorescence images overcomes the unreliability of atom trapping under low loading rates and avoids the influence of noise interference, thus improving the accuracy and reliability of atom identification in the corresponding scenarios. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 This is a flowchart of an automated recognition method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 2 This is a flowchart of a noise reduction process in an automated recognition method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the average fluorescence image in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a denoised image in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 5 This is a flowchart of a method for obtaining the average region radius in an automated recognition method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the fluorescence region segmentation in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 7 This is a flowchart of a method for obtaining the theoretical atomic positions of atoms in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of peak detection in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 9 This is a flowchart of the method for obtaining the optimal rotation angle in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 10 This is a schematic diagram illustrating the relationship between variance and rotation angle in an automated recognition method for atomic array fluorescence images provided by an embodiment of the present invention. Figure 11 This is a flowchart of a method for obtaining fluorescence intensity dataset in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 12 This is an example diagram illustrating the contribution weight calculation in an automated identification method for atomic array fluorescence images provided in this embodiment of the invention; Figure 13 This is a flowchart illustrating the determination of the presence of atoms in an automated identification method for atomic array fluorescence images provided in this embodiment of the invention. Figure 14 This is a schematic diagram illustrating the selection of intensity representative values ​​in an automated identification method for atomic array fluorescence images provided in an embodiment of the present invention; Figure 15 This is a schematic diagram illustrating the calculation of inter-class variance in an automated atomic array fluorescence image recognition method provided in an embodiment of the present invention; Figure 16 This is a schematic diagram of an automated recognition device for atomic array fluorescence images provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0022] Unless the context otherwise requires, throughout the specification and claims, the term "comprising" is interpreted as openly inclusive, meaning "including, but not limited to." In the description of the specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this disclosure. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics mentioned may be included in any suitable manner in any one or more embodiments or examples; that is, although they may be incorporated into embodiments or examples using the above terms for reasons such as order and position, it does not limit them to be incorporated in combination by a single embodiment or example.

[0023] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of this disclosure, unless otherwise stated, "a plurality of" means two or more. Furthermore, for example, the description may use the prefix "A" or "B" to describe the same type of nouns as two independent entities. In this case, the corresponding features defined with "A" and "B" are used only to distinguish between similar entities and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features.

[0024] In the description of this invention, the expression “A and / or B” (where A and B are used to formally represent specific features) will be used. The corresponding expression includes the following three combinations: only A, only B, and a combination of A and B.

[0025] As used in this invention, “about,” “approximately,” or “approximately” includes the stated value and the average value within an acceptable range of deviation from a particular value, wherein the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the particular quantity (i.e., the limitations of the measurement system).

[0026] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0027] Example 1: This embodiment provides an automated recognition method for atomic array fluorescence images, such as... Figure 1 As shown, the method flow includes the following.

[0028] In step 101, multiple atomic fluorescence images are acquired during atomic loading, and pixel averaging is performed to obtain an average fluorescence image.

[0029] This embodiment is applied in the scenario of loading and capturing atoms using an optical trap array. After collection and amplification, the atoms are detected by a camera to obtain an atomic fluorescence image of the captured state. In the image, the fluorescent dots arranged in the array represent the captured atoms. Because the loading rates of different capturing optical traps in the array are different, for the same atom, sometimes the atom can be successfully loaded, and sometimes it cannot. When the capturing optical trap successfully loads an atom, fluorescence will appear at the corresponding position in the atomic fluorescence image; when the capturing optical trap fails to load an atom, no fluorescence will appear at the corresponding position in the atomic fluorescence image. Considering the interference of external interference or astigmatism on the detection results in actual detection scenarios, and the loading rate of the capturing optical traps... Due to the instability of atomic fluorescence, the fluorescence in atomic fluorescence images may be difficult to distinguish, and it is difficult to accurately determine whether the trapping light trap at a certain position has successfully loaded atoms using a single atomic fluorescence image. Therefore, in this embodiment, multiple atomic captures are performed using a light trap array to obtain multiple atomic fluorescence images, and pixel averaging is performed to obtain an average fluorescence image. In the average fluorescence image, the brighter the fluorescence position (i.e., the larger the pixel value at the corresponding position), the greater the probability that an atom has been successfully captured at that position. The lower the fluorescence position (i.e., the smaller the pixel value at the corresponding position), the greater the probability that an atom has been successfully captured at that position. This is positively correlated with the loading rate of the trapping light trap at that position. Therefore, the average fluorescence image can better reflect whether an atom is loaded at the corresponding position.

[0030] In step 102, the average fluorescence image is denoised to obtain a denoised image. The region of the fluorescent pixels in the denoised image is segmented to obtain multiple fluorescent regions, and then the average region radius of the fluorescent region is obtained.

[0031] In this embodiment, it is first necessary to suppress background noise in the image to reduce noise interference on subsequent atomic fluorescence recognition. Therefore, the average fluorescence image needs to be denoised to obtain a denoised image. Furthermore, it is also necessary to determine the position of each fluorescent pixel in the denoised image and determine the fluorescence region based on the distribution of each fluorescent pixel. Each fluorescence region includes multiple adjacent fluorescent pixels, and each fluorescence region may correspond to a captured atom. The radius of the fluorescence region is obtained to determine the position and size of the fluorescence region, which facilitates the subsequent calculation of the fluorescence intensity of the fluorescence region to determine whether an atom is captured at the corresponding position.

[0032] In step 103, the theoretical atomic positions on the denoised image are obtained according to the type of the captured optical trap.

[0033] In this embodiment, the type of the light trap includes two types: The first method is applicable to optical trap arrays of any configuration, provided that the total number of traps is known and the fluorescence is clearly visible in the average fluorescence image, with the number of fluorescence points matching the number of traps. The second method is suitable for rectangular optical trap arrays, allowing for array rotation, and addresses situations where the fluorescence in the average fluorescence image is not sufficiently clear.

[0034] For the two different types of optical traps mentioned above, the theoretical atomic positions on the denoised image are calculated in a corresponding manner. The theoretical atomic positions are obtained from the fluorescent regions on the denoised image. It should be noted that the theoretical atomic positions are positions where atoms may be trapped, but not necessarily where atoms are present. Therefore, subsequent steps are needed to determine whether each theoretical atomic position actually contains trapped atoms.

[0035] In step 104, based on the average region radius and the theoretical atom position, a fluorescence intensity dataset of all atomic fluorescence images at the corresponding theoretical atom positions is obtained. Based on the fluorescence intensity dataset, a representative intensity value for the corresponding theoretical atom position is set, and the presence of an atom at the theoretical atom position is determined based on the average representative intensity value of all theoretical atom positions.

[0036] After obtaining the theoretical atomic positions, the fluorescence intensity of each atomic fluorescence image at the corresponding theoretical atomic position can be calculated to obtain a fluorescence intensity dataset. The intensity representative value is used to reflect the overall situation of whether the atom at the corresponding theoretical atomic position is captured. Based on the intensity representative value, it is finally determined whether there is a captured atom at the corresponding theoretical atomic position.

[0037] In this embodiment, an average fluorescence image is obtained by averaging the pixels of multiple atomic fluorescence images. Region segmentation is performed based on the differences in pixel values ​​on the average fluorescence image to determine the location and size of the fluorescence region where the atom may be located. This allows for the precise determination of the location on the average fluorescence image where the captured atom might be present. The presence of the atom at that location is determined by the fluorescence intensity of each atomic fluorescence image at that location. By superimposing and calculating multiple atomic fluorescence images, the unreliability of atom capture under low loading rates is overcome, and the influence of noise interference is avoided, thus improving the accuracy and reliability of atom identification in the corresponding scenario.

[0038] Furthermore, in this embodiment, in order to reduce the interference of background noise in the average fluorescence image on subsequent recognition, it is necessary to suppress the background noise in the image to reduce the misleading influence of pixel values ​​of pixels at non-light trap array locations in the background noise. Therefore, this embodiment also involves the following design: The average fluorescence image is denoised to obtain a denoised image, such as... Figure 2As shown, the method flow includes the following.

[0039] In step 201, the average fluorescence image is input as the input image into the initial iteration round.

[0040] Each iteration includes: In step 202, the arithmetic mean of the pixel values ​​of all pixels in the input image is calculated.

[0041] In this embodiment, the pixel value of all pixels in the input image is... The arithmetic mean is The pixel value can be the grayscale value of the corresponding pixel, used to represent the fluorescence intensity of the corresponding pixel. The larger the pixel value, the more likely the pixel is to be a pixel in the fluorescence region generated when the optical trap is loaded with atoms.

[0042] In step 203, an optimized image is generated, and pixels with pixel values ​​greater than 0 are designated as fluorescent pixels to obtain the number of fluorescent pixels in the optimized image.

[0043] In the optimized image, the pixel value of each pixel is... If the pixel value is greater than the arithmetic mean Then the pixel value of that pixel is converted to A value greater than 0 indicates that the pixel possesses sufficient brightness for optimized image rendering, i.e., it is a fluorescent pixel; if the pixel value is less than or equal to the arithmetic mean... Then the pixel value of that pixel is converted to A value less than or equal to 0 will not result in any brightness in the optimized image. Since pixels in background noise locations all possess some brightness, but it is typically low, it is highly likely to be less than the arithmetic mean. Therefore, through the above design, all pixel values ​​lower than the arithmetic mean are eliminated. This involves hiding most of the background noise pixels to achieve noise reduction.

[0044] In step 204, the optimized image is linearly normalized to a preset interval to obtain the output image.

[0045] The preset interval is set by those skilled in the art according to actual conditions. In this embodiment, the preset interval can be [0, 256]. The expression for the pixels in the output image is: ; in, To optimize the maximum pixel value in the image, To optimize the minimum pixel value in the image.

[0046] In step 205, it is determined whether the number of fluorescent pixels in this iteration meets the convergence condition.

[0047] In step 206, if the convergence condition is met, the output image is used as the denoised image.

[0048] In step 207, if the convergence condition is not met, the output image is used as the input image in the next iteration.

[0049] In this embodiment, the convergence condition is: ; in, Let be the number of fluorescent pixels in the k-th iteration. This represents the number of fluorescent pixels in the (k-1)th iteration. For the preset threshold, It can be 0.01.

[0050] like Figure 3 and Figure 4 As shown, Figure 3 The average fluorescence image, Figure 4 For the denoised image, it can be seen that Figure 4 Compared to Figure 3 Background noise was significantly removed.

[0051] Furthermore, in this embodiment, after acquiring the denoised image, it is necessary to identify all pixels (i.e., fluorescent pixels) corresponding to the capture light traps, and obtain the position and size of the fluorescent region corresponding to the capture light trap according to the distribution of the pixels. The corresponding design is as follows: the region of fluorescent pixels in the denoised image is divided into multiple fluorescent regions, and then the average region radius of each fluorescent region is obtained, such as... Figure 5 As shown, the method flow includes the following.

[0052] In step 301, the set of locations of all fluorescent pixels in the denoised image is obtained.

[0053] In this embodiment, the fluorescent pixel is a pixel with a non-zero pixel value, and the position set is the set of position coordinates of all fluorescent pixels.

[0054] In step 302, all adjacent fluorescent pixels are treated as a single fluorescent region.

[0055] Each time, a random fluorescent pixel is selected from the location set. A region growing algorithm is used to find all other fluorescent pixels image-connected to this pixel, thus obtaining the fluorescent region. Each fluorescent pixel in the fluorescent region is an adjacent pixel in the vertical, horizontal, and vertical directions. All found fluorescent pixels are removed from the location set. Another random location set is selected, and the above steps are repeated until all fluorescent pixels in the location set have been removed, resulting in all fluorescent regions. Figure 6 As shown, this is the denoised image after dividing the fluorescent region.

[0056] In step 303, the region radius of each fluorescent region in the denoised image is obtained, and then the average region radius corresponding to the denoised image is obtained.

[0057] In this embodiment, the expression for the region radius is: ; in, Let i be the average region radius of the fluorescent region. denoted as the number of fluorescent pixels in fluorescent region i.

[0058] The average region radius is the average of the region radii of all fluorescent regions in the denoised image.

[0059] Furthermore, in this embodiment, for different types of optical trap arrays, corresponding methods can be used to obtain the corresponding positions of the captured atoms on the denoised image. When the loading rate of the captured optical traps in the optical trap array meets the requirements, it means that the fluorescent area of ​​the captured optical trap on the denoised image is clearly identifiable, and the atomic capture rate is guaranteed. Therefore, the atom positions can be directly determined by the positions of each fluorescent pixel on the denoised image without image position adjustment, thus improving recognition efficiency. Therefore, this embodiment designs the following: when the atomic assembly rate of the captured optical trap is higher than or equal to a preset intensity representative value, obtaining the theoretical atom positions on the denoised image according to the type of captured optical trap specifically includes: The intensity centroid coordinates of each fluorescent region are calculated using the following expression: ; ; in, The x-coordinate of the intensity centroid coordinates The ordinate of the intensity centroid coordinates Let (x, y) be the set of all pixels contained in the i-th region, and (x, y) be the set of positions of these pixels. The image light intensity value of the pixel located at (x,y).

[0060] On the other hand, when the loading rate of the capture light traps in the optical trap array does not meet the requirements, it means that the clarity of the fluorescence region of the capture light traps in the denoised image is insufficient, and the capture rate of atoms cannot be guaranteed. In this case, for the optical trap array, not all capture light traps in the same row or column can be loaded with atoms. Therefore, when obtaining the atom positions on the denoised image, it is necessary to identify the positions with relatively higher pixel values ​​in each row and column, as these positions have a higher probability of containing atoms. Therefore, this case is suitable for rectangular optical trap arrays, and the optical trap array is allowed to have a rotation angle. The corresponding design is as follows: when the atom assembly rate of the capture light trap is higher than or equal to the preset intensity representative value, the theoretical atom positions on the denoised image are obtained according to the type of capture light trap, such as... Figure 7 As shown, the method flow includes: In step 401, the optimal rotation angle of the denoised image is obtained, and the denoised image is converted into an aligned image according to the optimal rotation angle. The pixel values ​​of all fluorescent pixels in the aligned image are then obtained.

[0061] In this embodiment, if the denoised image itself has a tilt angle, the optical trap array on the denoised image will also have the same tilt angle. The rows and columns on the optical trap array will not extend along the horizontal and vertical directions of the image. It will be impossible to project the rows onto the X-axis to sum the pixel values ​​of that row, or the columns onto the Y-axis to sum the pixel values ​​of that column. Furthermore, since the above scenario is for situations where the atomic loading rate of some atoms in a square array is insufficient, making it difficult to confirm the existence of the corresponding trap positions, it is necessary to indirectly obtain the positions of these missing traps by combining other row and column information. When the denoised image has a tilt angle, the corresponding row and column information needs to be extracted by selecting a suitable reference coordinate axis direction based on the tilt angle. Therefore, in this embodiment, it is necessary to first obtain the tilt angle of the denoised image, i.e., the optimal rotation angle, and rotate the denoised image according to the optimal rotation angle to become an aligned image. The rows in the optical trap array in the aligned image are parallel to the horizontal direction, and the columns in the optical trap array in the aligned image are parallel to the vertical direction. This facilitates the subsequent projection of the pixels of each row onto the X-axis to sum the pixel values ​​of that row, and also facilitates the subsequent projection of the pixels of each column onto the Y-axis to sum the pixel values ​​of that row. It should be noted that in this embodiment, when the denoised image is rotated, the denoised image may exceed the maximum range of the image. Therefore, at least one ring of zeros can be added around the original maximum range of the image to fill in the gaps and prevent the denoised image from exceeding the maximum range of the image after rotation.

[0062] In step 402, the row mean of pixel values ​​of each row in the aligned image is obtained. Local peak detection is performed on the row mean of all rows in the aligned image, and the row where the row mean is located at the local peak position is taken as the candidate row center. When there are other candidate row centers within the interval of the positive and negative average region radius in the vertical direction, the candidate row center with the highest row mean is retained and the other candidate row centers are removed.

[0063] In this embodiment, the row mean is the average pixel value of all fluorescent pixels in the corresponding row of the aligned image. There are differences in the row mean between different rows; the row with the larger the row mean, the greater the probability of atoms being present in that row. The candidate row center is the row where the row mean is located at a local peak position. The row mean on both sides of this row needs to satisfy a monotonically increasing left side and a monotonically decreasing right side, and the row mean of this row needs to be greater than the global pixel value mean of all pixels in the aligned image. The range of the positive and negative average region radius is... , wherein That is, the y-coordinate of the corresponding row. This is the average region radius. Since within the same fluorescent region, if an atom does exist, it will only be located at the position with the highest fluorescence intensity, when there are other candidate row centers within the interval of the positive and negative average region radii in the vertical direction, it means that multiple candidate row centers have been selected for the same fluorescent region. Therefore, it is necessary to select the candidate row center with the largest average value in the row direction to retain, and discard the other candidate row centers. If an atom exists in the corresponding fluorescent region, the retained candidate row center is the row where the atom is located.

[0064] In step 403, the column mean of pixel values ​​of each column of pixels in the aligned image is obtained. Local peak detection is performed on the column mean of all columns in the aligned image. The column with the column mean located at the local peak position is taken as the candidate column center. When there are other candidate column centers within the interval of the positive and negative average region radius in the horizontal direction, the candidate column center with the highest column mean is retained and the other candidate column centers are removed.

[0065] In this embodiment, the column mean is the average pixel value of all fluorescent pixels in the corresponding column of the aligned image. There are differences in the column mean between different columns; the larger the column mean, the greater the probability of atoms existing in the column. The candidate column center is the column where the column mean is located at a local peak position. The column mean on both sides of this column needs to satisfy a monotonically increasing left side and a monotonically decreasing right side, and the column mean of this column needs to be greater than the global pixel value mean of all pixels in the aligned image. The range of the positive and negative average region radius is... , wherein That is, the y-coordinate of the corresponding column. This refers to the average region radius. Since within the same fluorescent region, if an atom is indeed present, it will only be located at the position of highest fluorescence intensity, if other candidate column centers exist within the interval of the positive and negative average region radii in the vertical direction, it means that multiple candidate column centers have been selected for the same fluorescent region. Therefore, it is necessary to select the candidate column center with the largest column-wise average value to retain, and discard the other candidate column centers. If an atom exists in the corresponding fluorescent region, the retained candidate column center is the column where the atom is located. For example... Figure 8 This is a schematic diagram of peak detection.

[0066] In step 404, all candidate row centers and candidate column centers are combined to obtain the rotational atom positions in the aligned image. All rotational atom positions are then subjected to inverse coordinate transformation according to the optimal rotation angle to obtain the theoretical atom positions of all atoms in the denoised image.

[0067] The process of combining all candidate row centers and candidate column centers involves using the candidate row centers as the x-coordinates of the rotated atom positions and the candidate column centers as the column coordinates of the rotated atom positions. This combination yields all possible positions of the atoms. The following example illustrates this: In a 10×8 aligned image, there is a 2×2 optical trap array with candidate row centers at rows 7 and 4 and candidate column centers at columns 3 and 5. Therefore, all the rotating atom positions include (7,3), (4,3), (7,5) and (4,5).

[0068] Since the obtained rotational atom positions are based on the position coordinates on the aligned image after rotation, it is also necessary to restore the rotational atom positions to the corresponding coordinates on the denoised image according to the optimal rotation angle, which is the theoretical atom position.

[0069] Furthermore, regarding the optimal rotation angle, in this embodiment, it is necessary to traverse multiple rotation angles. Based on the image state after rotation, it is determined whether the current image is an aligned image. The corresponding method is as follows: The optimal rotation angle for obtaining the denoised image is as follows... Figure 9 As shown, the method flow includes: In step 501, within a preset angle range, the denoised image is rotated multiple times around the image center position according to a first preset angle step size, and a rotated image is obtained after each rotation.

[0070] In this embodiment, the preset angle range is set by those skilled in the art based on actual conditions, i.e., the initial rotation angle selection range. The preset angle range can be... The first preset angle step size is set by those skilled in the art according to the actual situation. In this embodiment, the first preset angle step size can be 0.5 degrees.

[0071] In this embodiment, after the denoised image is rotated, the coordinate transformation formula corresponding to the coordinates in the rotated image is: ; in,( , () represents the coordinates of the rotation center of the rotated image. The counterclockwise rotation angle from the denoised image to the rotated image, ( , ) represents the coordinates of the corresponding position in the rotated image. , ) represents the coordinates of the corresponding position in the rotated image in the denoised image.

[0072] In step 502, the coordinates of each fluorescent pixel in the rotated image in the denoised image are obtained as adjustment coordinates, the pixel values ​​of the four adjacent related pixels at the adjustment coordinates are obtained, and the pixel value of the corresponding fluorescent pixel in the rotated image is calculated based on the pixel values ​​of the related pixels.

[0073] In this embodiment, each fluorescent pixel in the original denoised image corresponds to a single pixel. However, after rotation, the angle of the denoised image changes, so each pixel also changes position and angle. The fluorescent pixel in the rotated denoised image will be divided by four adjacent pixels arranged in the top, bottom, left, and right directions near its original position. Therefore, if you want to obtain the pixel value of the corresponding fluorescent pixel in the rotated pixel, you need to calculate it based on the four adjacent pixels.

[0074] The expression for calculating the pixel value of the corresponding fluorescent pixel in the rotated image is: ; in, , , To adjust the x-coordinate of the coordinate system, To adjust the ordinate of the coordinate system, The associated pixel is located in the bottom left corner. The associated pixel is located in the bottom right corner. The associated pixel is located in the top left corner. The associated pixel is located in the upper right corner.

[0075] In step 503, based on the pixel values ​​of all fluorescent pixels in the rotated image, the row pixel mean of each row of pixels in the rotated image is obtained, and the column pixel mean of each column of pixels in the rotated image is obtained. Based on the row pixel mean of each row of pixels and the column pixel mean of each column of pixels, the variance of the rotated image is obtained.

[0076] In this embodiment, the expression for the sum of variances is: ; in, The sum of variances corresponding to the rotated image. The average value of the row pixels. The average value of the column pixels. The sign for variance is .

[0077] like Figure 10 The figure shows a schematic diagram illustrating the relationship between variance and rotation angle.

[0078] In step 504, the rotation angle with the largest variance within a preset angle range is obtained as the preferred angle. With the preferred angle as the center, multiple rotations are performed within a preset positive and negative interval with a second preset angle step size to obtain the variance of the rotation image for each rotation. The rotation angle with the largest variance is taken as the optimal rotation angle.

[0079] In this embodiment, after obtaining the preferred angle, considering that the first preset angle step size is relatively large, there may be relatively better rotation angles that have not been screened out, the system further traverses within a preset positive and negative interval near the preferred angle to find a better rotation angle. The preset positive and negative interval is set by those skilled in the art according to the actual situation, i.e., a positive and negative degree interval centered on the preferred angle. In this embodiment, the new iteration center is the best position of the previous iteration, and the positive and negative interval is the previous iteration step size. When the number of segments is N, the new iteration step size is 2 / N times the previous step size; for example... ,in For the preferred angle, the second preset angle step size is smaller than the first preset angle step size. The second preset angle step size can be: dividing 1° into 180 parts, i.e., 0.00278°. In this embodiment, the method described above for improving the accuracy of the rotation angle can be used to perform progressive scanning of the obtained rotation angle with smaller ranges and smaller steps multiple times, and finally obtain the optimal rotation angle, thereby significantly reducing the amount of calculation while ensuring the accuracy of angle positioning.

[0080] In this embodiment, after obtaining all theoretical atomic positions, it is necessary to further determine whether an atom actually exists at the corresponding theoretical atomic position. This needs to be determined based on the fluorescence intensity of the corresponding fluorescence region at the theoretical atomic position. Therefore, it is necessary to first obtain the fluorescence intensity of each atom's fluorescence image at the theoretical atomic position. Thus, this embodiment also involves the following design: The process involves obtaining a dataset of fluorescence intensity at the corresponding theoretical atomic positions for all atomic fluorescence images based on the average region radius and the theoretical atomic positions. Figure 11 As shown, it includes: In step 601, for each atomic fluorescence image, a square region with a side length of twice the average region radius is defined as the calculation region, centered on the theoretical atomic position.

[0081] The left boundary of the computational region is: .

[0082] The right boundary of the computational region is: .

[0083] The upper boundary of the computational region is: .

[0084] The lower boundary of the computational region is: .

[0085] in, The x-coordinate represents the theoretical atomic position. The ordinate represents the theoretical atomic position. The average radius of the region.

[0086] In step 602, all intersection pixels that intersect with the corresponding calculation region are obtained, and the overlap area between the intersection pixels and the calculation region is used as the contribution weight of the corresponding intersection pixels.

[0087] In this embodiment, the photosensitive area of ​​a single pixel can be considered as a unit square with a side length of 1. Therefore, for a pixel at coordinates (i, j), its photosensitive area is... If the pixel is a pixel at the intersection of the corresponding computational regions, the expression for calculating its contribution weight is:

[0088] in, This represents the overlap length between the intersecting pixels along the x-axis and the computational region. This represents the overlap length between the intersecting pixels along the y-axis and the computational region. The contribution weights for the intersection pixels (i, j).

[0089] To more intuitively illustrate the calculation method of the above contribution weights, such as Figure 12 For example, as can be seen from the image, the theoretical atomic position is located at the pixel. The center of the calculation area will be the pixel. Complete coverage, with each of the 8 surrounding pixels partially covered. The contribution weights of each pixel are as follows: Pixels It is completely covered, therefore the corresponding contribution weight 1; pixel , , and All were half-covered, therefore the corresponding contribution weights 0.5 pixels , , and Each was covered by one-quarter, therefore the corresponding contribution weight It is 0.25.

[0090] In step 603, the pixel values ​​of all intersecting pixels are summed according to their corresponding contribution weights to obtain the fluorescence feature intensity of the corresponding theoretical atomic position of the atomic fluorescence image.

[0091] The expression for the fluorescence characteristic intensity is: ; in, The fluorescence characteristic intensity, Let (i, j) be the pixel value at position (i, j). The contribution weight for position (i, j).

[0092] In step 604, the fluorescence feature intensities of all atomic fluorescence images at the corresponding theoretical atomic positions are obtained, which are used as the fluorescence intensity dataset at the corresponding theoretical atomic positions.

[0093] In this embodiment, the above steps are performed for each atomic fluorescence image, and the fluorescence feature intensity of a single theoretical atom position on all atomic fluorescence images is used as the fluorescence intensity dataset at the corresponding theoretical atom position.

[0094] Furthermore, the step involves setting a representative intensity value for the corresponding theoretical atomic position based on the fluorescence intensity dataset, and determining whether an atom exists at the theoretical atomic position based on the average of the representative intensity values ​​for all theoretical atomic positions. Figure 13 As shown, the specific steps include: In step 701, the intensity representative value corresponding to each theoretical atom position is obtained by the Otsu's method, and the preferred average fluorescence intensity value that is greater than the intensity representative value is obtained from the fluorescence intensity data corresponding to each theoretical atom position.

[0095] In this embodiment, obtaining the intensity representative value corresponding to each theoretical atomic position using the maximum inter-class variance method specifically involves: traversing all fluorescence feature intensities in the fluorescence intensity dataset as candidate thresholds; dividing the fluorescence intensity dataset into two parts using these candidate thresholds; forming one set of all fluorescence feature intensities greater than the candidate thresholds; and forming another set of all fluorescence feature intensities less than or equal to the candidate thresholds. }, }; in, For one of the partial sets, For another part of the set, Given a fluorescence intensity dataset, T represents the candidate threshold.

[0096] The corresponding inter-class variance is: ; in, For inter-class variance, for The mean of a partial set. for The mean of a partial set. for The probability corresponding to a partial set, for The probability corresponding to a partial set.

[0097] Each fluorescence feature intensity in the fluorescence intensity dataset is used as a candidate threshold, and the corresponding inter-class variance is calculated. The candidate threshold with the largest inter-class variance is then used as the representative intensity value. Figure 14 The diagram shown illustrates the selection of representative strength values; as shown... Figure 15 The diagram shown illustrates the calculation of inter-class variance.

[0098] The preferred average fluorescence intensity is... The mean of the intensities of all fluorescence features.

[0099] In step 702, the mean of the intensity representative values ​​for all theoretical atomic positions is calculated, i.e., the intensity representative mean.

[0100] In step 703, when the preferred fluorescence intensity average value corresponding to the corresponding theoretical atomic position is greater than or equal to the intensity representative average value, then there is an atom at the corresponding theoretical atomic position.

[0101] When the preferred fluorescence intensity average value corresponding to the corresponding theoretical atomic position is greater than or equal to the intensity representative average value, the signal representing the corresponding theoretical atomic position is strong, and it can be determined that an atom exists at that position; when the preferred fluorescence intensity average value corresponding to the corresponding theoretical atomic position is less than the intensity representative average value, the signal representing the corresponding theoretical atomic position is weak, and it can be determined that an atom does not exist at that position.

[0102] Example 2: like Figure 16 The diagram shown is a schematic representation of an automated atomic array fluorescence image recognition device according to an embodiment of the present invention. This automated atomic array fluorescence image recognition device includes one or more processors 41 and a memory 42.

[0103] Processor 41 and memory 42 can be connected via a bus or other means. Figure 16Taking the example of a connection between China and Israel via a bus.

[0104] The memory 42, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs and non-volatile computer-executable programs, such as the automated identification method for atomic array fluorescence images in the above embodiments. The processor 41 executes the automated identification method for atomic array fluorescence images by running the non-volatile software programs and instructions stored in the memory 42.

[0105] Memory 42 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 42 may optionally include memory remotely located relative to processor 41, which can be connected to processor 41 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0106] The program instructions / modules are stored in the memory 42 and, when executed by one or more processors 41, perform the automated identification method for atomic array fluorescence images in the above embodiments.

[0107] This invention also provides a computer storage medium storing computer program instructions; when executed by a processor, the computer program instructions implement the automated identification method for atomic array fluorescence images provided in this invention.

[0108] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automated recognition method for atomic array fluorescence images, characterized in that, include: Multiple atomic fluorescence images were acquired during atomic loading, and pixel averaging was performed to obtain an average fluorescence image. The average fluorescence image is denoised to obtain a denoised image. The region of fluorescent pixels in the denoised image is segmented to obtain multiple fluorescent regions, and the average region radius of the fluorescent regions is obtained. Based on the type of optical trap, obtain the theoretical atomic positions on the denoised image; Based on the average region radius and the theoretical atom position, a fluorescence intensity dataset of all atomic fluorescence images at the corresponding theoretical atom positions is obtained. Based on the fluorescence intensity dataset, a representative value of the intensity at the corresponding theoretical atom position is set, and the presence of an atom at the theoretical atom position is determined based on the average of the representative values ​​of the intensity at all theoretical atom positions.

2. The automated recognition method for atomic array fluorescence images according to claim 1, characterized in that, The step of denoising the average fluorescence image to obtain a denoised image specifically includes: The average fluorescence image is used as the input image in the initial iteration round, and each iteration round includes: Calculate the pixel value of all pixels in the input image. Arithmetic mean ; Generate an optimized image, optimizing the pixel value of each pixel in the image. Pixels with a value greater than 0 are identified as fluorescent pixels, and the number of fluorescent pixels in the optimized image is obtained. The optimized image is linearly normalized to a preset interval to obtain the output image; Determine whether the number of fluorescent pixels in this iteration satisfies the convergence condition; If the convergence condition is met, the output image is used as the denoised image; If the convergence condition is not met, the output image is used as the input image in the next iteration.

3. The automated recognition method for atomic array fluorescence images according to claim 2, characterized in that, The step of segmenting the region of fluorescent pixels in the denoised image to obtain multiple fluorescent regions and obtaining the average region radius of the fluorescent regions specifically includes: Obtain the set of locations of all fluorescent pixels in the denoised image; Treat all adjacent fluorescent pixels as a single fluorescent region; Obtain the region radius of each fluorescent region in the denoised image, and then obtain the average region radius of the denoised image.

4. The automated recognition method for atomic array fluorescence images according to claim 1, characterized in that, When the atomic assembly rate of the captured optical trap is higher than or equal to a preset intensity representative value, the step of obtaining the theoretical atomic positions on the denoised image according to the type of captured optical trap specifically includes: The intensity centroid coordinates of each fluorescent region are calculated using the following expression: ； ； in, The x-coordinate of the intensity centroid coordinates The ordinate of the intensity centroid coordinates Let i be the set of all pixels contained in the i-th region. The image light intensity value of the pixel located at (x,y).

5. The automated recognition method for atomic array fluorescence images according to claim 1, characterized in that, When the atomic assembly rate of the captured optical trap is lower than a preset intensity representative value, the step of obtaining the theoretical atomic positions on the denoised image according to the type of captured optical trap specifically includes: Obtain the optimal rotation angle of the denoised image, convert the denoised image into an aligned image based on the optimal rotation angle, and obtain the pixel values ​​of all fluorescent pixels in the aligned image; Obtain the row mean of pixel values ​​for each row in the aligned image. Perform local peak detection on the row mean of all rows in the aligned image and select the row where the row mean is located at the local peak position as the candidate row center. When there are other candidate row centers within the interval of the positive and negative average regions in the vertical direction, retain the candidate row center with the highest row mean and remove the other candidate row centers. Obtain the column mean of pixel values ​​in each column of the aligned image. Perform local peak detection on the column mean of all columns in the aligned image. Columns whose column mean is located at the local peak position are selected as candidate column centers. When there are other candidate column centers within the interval of the positive and negative average regions in the horizontal direction, retain the candidate column center with the highest column mean and remove the other candidate column centers. By combining all candidate row centers and candidate column centers, the rotational atom positions in the aligned image are obtained. All rotational atom positions are then subjected to inverse coordinate transformation according to the optimal rotation angle to obtain the theoretical atom positions of all atoms in the denoised image.

6. The automated recognition method for atomic array fluorescence images according to claim 5, characterized in that, The process of obtaining the optimal rotation angle for the denoised image specifically includes: Within a preset angle range, the denoised image is rotated multiple times around the center position of the image according to the first preset angle step size, and a rotated image is obtained after each rotation; The coordinates of each fluorescent pixel in the rotated image in the denoised image are obtained as adjustment coordinates. The pixel values ​​of the four adjacent related pixels at the adjustment coordinates are obtained. The pixel value of the corresponding fluorescent pixel in the rotated image is calculated based on the pixel values ​​of the related pixels. Based on the pixel values ​​of all fluorescent pixels in the rotated image, obtain the row pixel mean of each row of pixels in the rotated image, obtain the column pixel mean of each column of pixels in the rotated image, and obtain the variance sum of the rotated image based on the row pixel mean of each row of pixels and the column pixel mean of each column of pixels. The rotation angle with the largest variance within a preset angle range is selected as the preferred angle. With the preferred angle as the center, multiple rotations are performed within a preset positive and negative interval with a second preset angle step size. The variance of the rotation image for each rotation is obtained, and the rotation angle with the largest variance is selected as the optimal rotation angle.

7. The automated recognition method for atomic array fluorescence images according to claim 6, characterized in that, The step of calculating the pixel value of the corresponding fluorescent pixel in the rotated image based on the pixel value of the associated pixel specifically includes: The expression for calculating the pixel value of the corresponding fluorescent pixel in the rotated image is: ； in, , , To adjust the x-coordinate of the coordinate system, To adjust the ordinate of the coordinate system, The associated pixel is located in the bottom left corner. The associated pixel is located in the bottom right corner. The associated pixel is located in the top left corner. The associated pixel is located in the upper right corner.

8. The automated recognition method for atomic array fluorescence images according to claim 6, characterized in that, The step of obtaining the fluorescence intensity dataset of all atomic fluorescence images at the corresponding theoretical atomic positions based on the average region radius and theoretical atomic positions specifically includes: For each atomic fluorescence image, a square region with a side length of twice the average region radius is defined as the calculation region, centered on the theoretical atom position; Obtain all intersection pixels that intersect with the corresponding calculation region, and use the overlap area between the intersection pixels and the calculation region as the contribution weight of the corresponding intersection pixels; The pixel values ​​of all intersecting pixels are weighted and summed according to their respective contribution weights to obtain the fluorescence feature intensity of the corresponding theoretical atomic position in the atomic fluorescence image. The fluorescence characteristic intensities of all atomic fluorescence images at the corresponding theoretical atomic positions are obtained and used as a fluorescence intensity dataset at the corresponding theoretical atomic positions.

9. An automated recognition device for atomic array fluorescence images, characterized in that, The method includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor for performing the automated identification method for atomic array fluorescence images according to any one of claims 1-8.

10. A non-volatile computer storage medium, characterized in that, The computer storage medium stores computer program instructions that, when executed by one or more processors, implement the automated identification method for atomic array fluorescence images as described in any one of claims 1-8.

Images