Photoetching mask optimization method based on improved particle swarm optimization algorithm

Abstract

The invention discloses a photoetching mask optimization method based on an improved particle swarm optimization algorithm. The photoetching mask optimization method comprises the following steps: acquiring a photoetching mask image and a target image; performing Hopkin photoetching simulation on the photoetching mask image to obtain a simulation image; calculating a deviation between the simulation image and a target image, and obtaining an image deviation; based on the image deviation, utilizing an improved particle swarm optimization method to optimize a photoetching mask image to obtain an optimal mask, and introducing a speed pause strategy, dividing double populations and adding a crossover variation mechanism into the particle swarm optimization method to obtain the improved swarm optimization method. A speed pause strategy is introduced, a new choice is provided for speed updating of particles, and the ability of the algorithm to get rid of premature convergence is greatly improved. By dividing the two subgroups and adding crossover and mutation operation in the second subgroup, the diversity of the population is ensured, and the global optimization ability of the algorithm is significantly enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of microelectronics manufacturing, and in particular relates to a photolithography mask optimization method based on an improved particle swarm optimization algorithm. Background Art

[0002] In microelectronics manufacturing, photolithography, as a crucial step, remains a key driver of continuous advancement. Masks are a core component of the lithography process, and their design and optimization directly impact the final image quality. However, as manufacturing processes advance toward smaller feature sizes, effective photolithography mask optimization methods are urgently needed to meet higher precision requirements.

[0003] Particle swarm optimization (PSO) algorithms are widely used in the field of photolithography due to their simple structure and fast convergence. In mask optimization, a predefined fitness function is used. Through an iterative optimization process, the algorithm continuously adjusts the mask's design parameters to reduce the fitness function, thereby achieving quantitative evaluation and continuous improvement of image quality. However, as the dimensionality and complexity of problems continue to increase, the global search capabilities of traditional PSO algorithms may be limited, making them prone to local optima and premature convergence in the vast solution space. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a photolithography mask optimization method based on an improved particle swarm optimization algorithm, which can improve the final photolithography imaging quality.

[0005] To achieve the above objectives, the present invention provides a photolithography mask optimization method based on an improved particle swarm optimization algorithm, comprising:

[0006] acquiring a photolithography mask image and a target image;

[0007] Performing Hopkins lithography simulation on the photolithography mask image to obtain a simulation image;

[0008] Calculating the deviation between the simulated image and the target image to obtain the image deviation;

[0009] Based on the image deviation, the photolithography mask image is optimized using an improved particle swarm optimization method to obtain the optimal mask, wherein a speed pause strategy is introduced into the particle swarm optimization method, dual populations are divided, and a crossover mutation mechanism is added to obtain the improved swarm optimization method.

[0010] Optionally, performing Hopkins lithography simulation on the lithography mask image to obtain a simulation image includes:

[0011] Calculating the spatial frequency coordinates of the photolithography mask image;

[0012] Obtaining a cross-transmission coefficient based on the spatial frequency coordinates;

[0013] The photolithography mask image is simulated based on the cross transmission coefficient to obtain a simulation image.

[0014] Optionally, obtaining a cross-transmission coefficient based on the spatial frequency coordinate includes:

[0015] Based on the spatial frequency coordinates, performing element-by-element multiplication on the light source function and the impulse response function to obtain two sets of data;

[0016] Perform a convolution integral operation on the two sets of data to obtain the cross-transmission coefficient.

[0017] Optionally, simulating the photolithography mask image based on the cross-transmission coefficient to obtain a simulated image includes:

[0018] performing a Fourier transform on the photolithography mask image based on the cross-transfer coefficient;

[0019] Based on the transformed lithography mask image and combined with the lithography simulation function, the electric field distribution is obtained;

[0020] Perform inverse Fourier transform on the electric field distribution to obtain the simulation image.

[0021] Optionally, calculating the deviation between the simulated image and the target image to obtain the image deviation includes:

[0022] Binarizing the simulated image to obtain a photolithography binary image;

[0023] The deviation between the lithography binary image and the target image is calculated to obtain the image deviation.

[0024] Optionally, obtain improved particle swarm optimization methods including:

[0025] Randomly initialize a population, wherein each particle in the population represents a mask solution;

[0026] Improve the speed update formula based on the initialized population;

[0027] Based on the improved speed update formula, the initialized population is divided into two subgroups, one for local development and the other for global search;

[0028] Based on local search subgroups, crossover and mutation operations are embedded in the subgroups;

[0029] Perform wide-area exploration based on global search subgroups;

[0030] Based on the local search subgroup and the global search subgroup, an improved particle swarm optimization method is obtained.

[0031] Optionally, improving the speed update formula includes incorporating a speed pause mechanism.

[0032] Optionally, optimizing the photolithography mask image based on the image deviation using an improved particle swarm optimization method to obtain the optimal mask includes:

[0033] Calculate image error based on improved particle swarm optimization algorithm;

[0034] randomly generating an initial mask solution based on the image deviation;

[0035] Setting particle swarm optimization parameters based on the initial mask solution;

[0036] Iteration is performed based on the particle swarm optimization parameters to obtain the optimal mask.

[0037] Optionally, iterating based on the particle swarm optimization parameters to obtain the optimal mask includes:

[0038] Calculate the fitness value of the particle;

[0039] Based on the fitness value, updating the individual optimal solution and the global optimal solution;

[0040] Based on the individual optimal solution and the global optimal solution, updating the optimized mask and the deviation value;

[0041] The optimal mask is obtained based on the optimized mask and the bias value.

[0042] Compared with the prior art, the present invention has the following advantages and technical effects:

[0043] This invention introduces a velocity pause strategy, providing a new option for particle velocity updates and significantly improving the algorithm's ability to avoid premature convergence. By dividing the population into two subgroups and adding a crossover mutation operation to the second subgroup, this ensures population diversity and significantly enhances the algorithm's global optimization capabilities. This invention uses multiple metrics, including graph deviation and graph similarity, to comprehensively evaluate exposure imaging accuracy, validating the optimization algorithm's superior performance and practical application in improving imaging quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0045] Figure 1 This is a flow chart of a photolithography mask optimization method based on an improved particle swarm optimization algorithm according to an embodiment of the present invention;

[0046] Figure 2is a schematic diagram of an initial mask image according to an embodiment of the present invention;

[0047] Figure 3 is a schematic diagram of a digital lithography system according to an embodiment of the present invention;

[0048] Figure 4 is a schematic diagram of a crossover mutation operation according to an embodiment of the present invention;

[0049] Figure 5 Schematic diagram of exposure images before and after optimization according to an embodiment of the present invention;

[0050] Figure 6 1 is a schematic diagram of mask images before and after optimization according to an embodiment of the present invention;

[0051] Figure 7 1 is a graph of convergence curves obtained for different mask patterns in an embodiment of the present invention, wherein (a) is a graph of convergence curve for pattern one, (b) is a graph of convergence curve for pattern two, and (c) is a graph of convergence curve for pattern three;

[0052] Figure 8 This is a graph of the algorithm calculation time consumption of an embodiment of the present invention. DETAILED DESCRIPTION

[0053] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0054] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0055] This embodiment provides a photolithography mask optimization method based on an improved particle swarm optimization algorithm. Figure 1 As shown, the specific steps include:

[0056] acquiring a photolithography mask image and a target image;

[0057] Performing Hopkins lithography simulation on the lithography mask image to obtain a simulation image;

[0058] Calculate the deviation between the simulated image and the target image to obtain the image deviation;

[0059] Based on image deviation, the improved particle swarm optimization method is used to optimize the photolithography mask image to obtain the optimal mask. A speed pause strategy is introduced into the particle swarm optimization method, and a dual population is divided and a crossover mutation mechanism is added to obtain an improved swarm optimization method.

[0060] Specifically, this embodiment includes: S1, importing a pre-designed photolithography mask image and a target image;

[0061] S2. Define the light source function and impulse response function;

[0062] S3. Define a cross transmission coefficient and calculate the coefficient (TCC);

[0063] S4. Define the transfer function;

[0064] S5, performing Hopkins lithography simulation on the imported initial mask image to obtain an image after lithography simulation;

[0065] S6. Define a deviation function PE to calculate the deviation between the simulated image and the target image;

[0066] S7, generating an initial solution and optimizing the mask image using a particle swarm optimization algorithm;

[0067] S8, introduce the velocity pause strategy to provide a third option for particle velocity update;

[0068] S9, divide the population into two groups, and introduce the crossover and mutation mechanism of the genetic algorithm into the second group;

[0069] S10, applying the improved particle swarm optimization algorithm to iteratively optimize the mask;

[0070] S11, output the optimized optimal mask and pattern deviation PE;

[0071] S12. Visualization results, including the transformation curve of the fitness value and the computational efficiency graph of the algorithm.

[0072] Furthermore, performing Hopkins lithography simulation on the lithography mask image to obtain a simulation image includes:

[0073] Calculate the spatial frequency coordinates of the lithography mask image;

[0074] Based on the spatial frequency coordinates, a cross transmission coefficient is obtained;

[0075] The photolithography mask image is simulated based on the cross transmission coefficient to obtain a simulation image.

[0076] Furthermore, based on the spatial frequency coordinates, obtaining the cross-transmission coefficient includes:

[0077] Based on the spatial frequency coordinates, the light source function and the impulse response function are multiplied element by element to obtain two sets of data;

[0078] Perform convolution integration operation on the two sets of data to obtain the cross-transmission coefficient.

[0079] Specifically, the spatial frequency coordinates are calculated;

[0080] Multiply the light source function and the impulse response function element by element to obtain two sets of data;

[0081] Perform convolution integration operation on the two sets of data.

[0082] More specifically, the light source function and impulse response function are:

[0083]

[0084] Where σ is the coherence of the exposure light source, λ is the wavelength of the light source, and NA is the numerical aperture.

[0085] The transfer function is:

[0086]

[0087] Among them, z is the distance from the light source to the object, f x and f y Frequency coordinate, n is the refractive index, and j is the imaginary unit.

[0088] Furthermore, simulating the photolithography mask image based on the cross-transmission coefficient, and obtaining the simulated image includes:

[0089] Performing Fourier transform on the lithography mask image based on the cross-transfer coefficient;

[0090] Based on the transformed lithography mask image and combined with the lithography simulation function, the electric field distribution is obtained;

[0091] Perform inverse Fourier transform on the electric field distribution to obtain a simulation image.

[0092] Specifically, the imported initial mask image is subjected to Hopkins lithography simulation to obtain the image after lithography simulation, including

[0093] Calculate the Fourier transform of the initial mask;

[0094] Apply various lithography simulation functions to obtain the electric field distribution E;

[0095] Perform inverse Fourier transform on the electric field E to obtain a simulation image after photolithography simulation.

[0096] Furthermore, the deviation between the simulated image and the target image is calculated. Obtaining the image deviation includes:

[0097] Binarize the simulation image to obtain a photolithography binary image;

[0098] The deviation between the lithography binary image and the target image is calculated to obtain the image deviation.

[0099] Specifically, define the deviation function PE to calculate the deviation between the simulated image and the target image, including:

[0100] Define a hard threshold function;

[0101] Binarizing the simulated image after the lithography simulation to obtain a lithography binary image;

[0102] Calculate the difference between the lithography binary image and the target image, take the square, and sum them to obtain the graphic deviation PE.

[0103] Furthermore, the improved particle swarm optimization method is obtained including:

[0104] Randomly initialize the population, where each particle in the population represents a mask solution;

[0105] Improve the speed update formula based on the initialized population;

[0106] Based on the improved speed update formula, the initialized population is divided into two subgroups, one for local development and the other for global search;

[0107] Based on local search subgroups, crossover and mutation operations are embedded in the subgroups;

[0108] Perform wide-area exploration based on global search subgroups;

[0109] Based on the local search subgroup and the global search subgroup, an improved particle swarm optimization method is obtained.

[0110] Furthermore, improving the speed update formula includes incorporating a speed pause mechanism.

[0111] Furthermore, based on the image deviation, the improved particle swarm optimization method is used to optimize the lithography mask image to obtain the optimal mask, including:

[0112] Calculate image error based on improved particle swarm optimization algorithm;

[0113] Based on the image deviation, the initial mask solution is randomly generated;

[0114] Based on the initial mask solution, set the particle swarm optimization parameters;

[0115] Iterate based on the particle swarm optimization parameters to obtain the optimal mask.

[0116] Furthermore, the optimal mask is obtained by iterating based on the particle swarm optimization parameters, including:

[0117] Calculate the fitness value of the particle;

[0118] Based on the fitness value, update the individual optimal solution and the global optimal solution;

[0119] Based on the individual optimal solution and the global optimal solution, update the optimized mask and deviation value;

[0120] Based on the optimized mask and bias value, the optimal mask is obtained.

[0121] Specifically, the particle swarm optimization algorithm is used to optimize the mask image, including:

[0122] S71, randomly generating an initial mask solution;

[0123] S72, setting particle swarm optimization algorithm parameters;

[0124] The population size is 50, the individual learning factor is 1.2, the social learning factor is 0.3, the inertia weight is 0.9, and the number of iterations is 5000;

[0125] S73, iterative process of particle swarm optimization algorithm;

[0126] Calculate the fitness value of the particle;

[0127] Update individual optimal solutions and global optimal solutions;

[0128] Update the optimized mask and the deviation value PE;

[0129] As a further improvement to this application, a velocity pause strategy is introduced to provide a third option for particle velocity updates, including:

[0130] Modify the speed update formula;

[0131] Introduced speed pause control parameters to dynamically adjust particle update options;

[0132] As a further improvement of the present application, two populations are divided, and the second population introduces the crossover and mutation mechanism of the genetic algorithm, including:

[0133] Split the original population into two, i.e. the sizes of the two new sub-populations are 25 respectively;

[0134] The second population adds the crossover and mutation operations of the genetic algorithm, with the crossover probability and mutation probability being 0.8 and 0.2 respectively, to maintain the diversity of the population;

[0135] As a further improvement of this application, the optimized optimal mask and pattern deviation PE are output, including:

[0136] Iterative optimization based on improved particle swarm optimization algorithm;

[0137] Output graph minimum deviation value PE;

[0138] Output the corresponding optimized mask.

[0139] The present embodiment will be described in detail below with reference to the accompanying drawings:

[0140] This embodiment provides a method for optimizing a photolithography mask based on an improved particle swarm optimization algorithm. Figure 1 As shown, including:

[0141] S1. Import pre-designed photolithography mask image and target image, such as Figure 2 As shown in the figure, the letters Y and B and irregular shapes are selected as three different masks in this experiment, and the size of each mask is 30*30 pixels;

[0142] S2. Define the light source function and impulse response function; the light source function and impulse response function are:

[0143]

[0144] Where σ is the coherence of the exposure light source, λ is the wavelength of the light source set to 405 nm, and NA is the numerical aperture set to 0.5.

[0145] S3. Define the cross transmission coefficient and calculate the coefficient (TCC); the formula is described as:

[0146] TCC(f',g',f”,g”)=∫∫J(f,g)P(f+f’,g+g’)P*(f+f”,g+g”)dfdg,

[0147] S4. Define the transfer function; the transfer function is:

[0148]

[0149] Among them, z is the distance from the light source to the object, which is set to 0.8 meters, f x and f y Frequency coordinate, n is the refractive index set to 1.5, and j is the imaginary unit.

[0150] Specifically, Figure 3 A digital lithography system is demonstrated, using a lithography imaging system model based on the Hopkins model. A 405-nanometer ultraviolet light-emitting diode (UVLED) is used as the light source, a dynamic dynamic range detector (DMD) dynamically adjusts the digital mask, and a three-dimensional motion platform precisely controls the depth of focus. In this model, the image intensity distribution calculation involves key parameters such as spatial coordinates, frequency coordinates, complex conjugate, spatial light intensity, inverse Fourier transform, Fourier transform of the mask, cross-intensity function, coherence, numerical aperture, refractive index, and pupil function. The following formula accurately describes the image intensity within the lithography field.

[0151] I(x,y)=F -1 {∫∫∫∫TCC(f',g',f”,g”)M(f',g')M *(f”,g”)df'dg'df”dg”},

[0152] S5, performing Hopkins lithography simulation on the imported initial mask image to obtain an image after lithography simulation;

[0153] S6. Define the deviation function PE to calculate the deviation between the simulated image and the target image; the formula is as follows:

[0154]

[0155] S7. Generate an initial solution and use the particle swarm optimization algorithm to optimize the mask image; the population size is set to 50, the individual learning factor and the social learning factor are 1.2 and 0.3 respectively, the inertia weight is set to 0.9, and the number of iterations is set to 5000.

[0156] S8. Introducing the velocity pause strategy provides a third option for particle velocity updates. The specific implementation process can be described by the following formula:

[0157]

[0158] a is a velocity pause parameter, and if its value is greater than 1, all particles will update their velocity in each iteration in exactly the same way as the traditional PSO algorithm. On the other hand, extremely low values will force particles to move at a constant velocity and restrict them from moving faster or slower.

[0159] S9, divide the population into two groups, the size of each subgroup is 25, and the crossover and mutation mechanism of the genetic algorithm is introduced into the second subgroup; Figure 4 As shown in FIG, the specific operations of the crossover process and the mutation process are shown, where the crossover probability and the mutation probability are 0.8 and 0.2 respectively.

[0160] S10, applying the improved particle swarm optimization algorithm to iteratively optimize the mask;

[0161] S11, output the optimized optimal mask and graphic deviation PE; and introduce the structural similarity index SSIM and the binary error R of the mask BE and the total change of the mask R TV As an evaluation indicator of algorithm performance, the specific formula is as follows:

[0162]

[0163] Among them, μ x ,μ y represents the average value of the difference between the exposed image and the target image, σ x ,σ y Represents the variance of the difference between the exposed image and the target image, σ xyRepresents the covariance of the difference between the exposed image and the target image. The result is as follows Figure 5 、 Figure 6 shown.

[0164] Before optimization, the PE values of the post-exposure images obtained from the mask were 31, 27, and 73, respectively, and the SSIM values were 0.80, 0.84, and 0.75, respectively. After optimizing the lithography mask using the particle swarm optimization algorithm, the PE values were 11, 13, and 30, respectively, and the SSIM values were 0.94, 0.93, and 0.90, respectively. The PE values were reduced by 51.8%-64.5%, and the SSIM values were increased to above 0.90. After optimizing the lithography mask using the improved particle swarm optimization algorithm, the PE values were 6, 4, and 18, respectively, and the SSIM values were 0.98, 0.96, and 0.95, respectively. Compared to the traditional particle swarm optimization algorithm, the PE values were reduced by 40.0%-69.2%, and the SSIM values were increased to above 0.95.

[0165] Binary error R of the photolithography mask after optimization based on particle swarm optimization algorithm BE They are 46, 14, and 89 respectively, with a total change of R TV The binary error R of the photolithography mask after optimization based on the improved particle swarm optimization algorithm is 0.65, 0.20, and 1.26 respectively. BE They are 54, 34, and 93 respectively, with a total change of R TV The values of 0.72, 0.50 and 1.29 respectively ensure the feasibility and reliability of the optimized mask in practical applications.

[0166] S12, visualization results, including the transformation curve of fitness value and the calculation efficiency graph of the algorithm. Figure 7 (a)-(c), Figure 8As shown in the figure, this study selected X / Y axisymmetric letters (B, Y) and irregular semiconductor line patterns with mixed right-angle / rounded corner features to ensure excellent geometric feature diversity and contrast. During the iterative process of mask optimization, both the traditional particle swarm optimization algorithm and the improved particle swarm optimization algorithm demonstrated effectiveness in improving lithography quality. The convergence curve shows that the improved particle swarm optimization algorithm achieved the highest accuracy. The computational cycle times of the traditional particle swarm optimization algorithm were 28.62s, 29.36s, and 28.28s, respectively, while the computational cycle times of the speed-pause-based particle swarm optimization algorithm were 208.99s, 209.45s, and 212.66s, respectively. Although this strategy makes up for the shortcoming of traditional methods that are prone to falling into local optimality in high-dimensional problems, it sacrifices a lot in terms of efficiency. On this basis, the improved particle swarm optimization algorithm that incorporates the crossover and mutation operations of the genetic algorithm has a calculation cycle of 27.21s, 27.71s, and 27.26s respectively. While greatly improving the accuracy, the calculation efficiency is also guaranteed to a certain extent, showing great potential in the field of micro-electronic manufacturing.

[0167] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A photolithography mask optimization method based on an improved particle swarm optimization algorithm, characterized in that: include: acquiring a photolithography mask image and a target image; Performing Hopkins lithography simulation on the photolithography mask image to obtain a simulation image; Calculating the deviation between the simulated image and the target image to obtain the image deviation; Based on the image deviation, the photolithography mask image is optimized using an improved particle swarm optimization method to obtain the optimal mask, wherein a speed pause strategy is introduced into the particle swarm optimization method, dual populations are divided, and a crossover mutation mechanism is added to obtain the improved swarm optimization method.

2. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that: Performing Hopkins lithography simulation on the lithography mask image to obtain a simulation image includes: Calculating the spatial frequency coordinates of the photolithography mask image; Obtaining a cross-transmission coefficient based on the spatial frequency coordinates; The photolithography mask image is simulated based on the cross transmission coefficient to obtain a simulation image.

3. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 2, characterized in that: Based on the spatial frequency coordinates, obtaining a cross-transmission coefficient includes: Based on the spatial frequency coordinates, performing element-by-element multiplication on the light source function and the impulse response function to obtain two sets of data; A convolution integral operation is performed on the two sets of data to obtain the cross-transmission coefficient.

4. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 2, characterized in that: Simulating the photolithography mask image based on the cross transmission coefficient, and obtaining a simulated image includes: performing a Fourier transform on the photolithography mask image based on the cross-transfer coefficient; Based on the transformed lithography mask image and combined with the lithography simulation function, the electric field distribution is obtained; Perform inverse Fourier transform on the electric field distribution to obtain the simulation image.

5. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that: Calculating the deviation between the simulated image and the target image, and obtaining the image deviation includes: Binarizing the simulated image to obtain a photolithography binary image; The deviation between the lithography binary image and the target image is calculated to obtain the image deviation.

6. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 1, characterized in that: Improved particle swarm optimization methods include: Randomly initialize a population, wherein each particle in the population represents a mask solution; Improve the speed update formula based on the initialized population; Based on the improved speed update formula, the initialized population is divided into two subgroups, one for local development and the other for global search; Based on local search subgroups, crossover and mutation operations are embedded in the subgroups; Perform wide-area exploration based on global search subgroups; Based on the local search subgroup and the global search subgroup, an improved particle swarm optimization method is obtained.

7. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 6, characterized in that: Improvements to the speed update formula include: incorporating a speed pause mechanism.

8. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 6, characterized in that: Based on the image deviation, the improved particle swarm optimization method is used to optimize the photolithography mask image to obtain the optimal mask, including: Calculate image error based on improved particle swarm optimization algorithm; randomly generating an initial mask solution based on the image deviation; Setting particle swarm optimization parameters based on the initial mask solution; Iteration is performed based on the particle swarm optimization parameters to obtain the optimal mask.

9. The photolithography mask optimization method based on the improved particle swarm optimization algorithm according to claim 8, characterized in that: Iterating based on the particle swarm optimization parameters to obtain the optimal mask includes: Calculate the fitness value of the particle; Based on the fitness value, updating the individual optimal solution and the global optimal solution; Based on the individual optimal solution and the global optimal solution, updating the optimized mask and the deviation value; The optimal mask is obtained based on the optimized mask and the bias value.

Images