Sample generation method and device and storage medium

Abstract

The invention discloses a sample generation method and device and a storage medium, and relates to the technical field of computational chemistry. The method comprises the following steps: determining at least one predicted reaction product of a selected reactant; for each predicted reaction product, generating a surface topography simulation diagram of the predicted reaction product based on the molecular structure of the predicted reaction product; and on the basis of the at least one surface topography simulation diagram of the predicted reaction product, generating a product simulation diagram as a training sample of a material analysis model, so that sample acquisition does not depend on real data, a large amount of real experimental data does not need to be manually labeled, the sample acquisition efficiency is improved, and the labor cost is reduced.

Description

Technical Field

[0001] The present application relates to the field of computational chemistry technology, and in particular to a sample generation method, device and storage medium. Background Art

[0002] In the related technology, experimental image analysis of surface materials mainly relies on machine learning models, which usually require a large amount of labeled data for training. The labeled data is generally performed manually by staff, which is not only time-consuming and labor-intensive, but also costly. In many cases, in order to obtain more detailed analysis results, very fine annotation is required, which is very difficult in actual operation.

[0003] Therefore, how to obtain model training samples more conveniently and quickly is a problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of this application is to provide a sample generation method, device and storage medium, aiming to solve the technical problem of how to obtain model training samples more conveniently and quickly.

[0005] To achieve the above objectives, the present application proposes a sample generation method, which includes: determining at least one predicted reaction product of the selected reactant; the predicted reaction product being an organic material; For each of the predicted reaction products, based on the molecular structure of the predicted reaction product, generating a surface morphology simulation image of the predicted reaction product; Based on the surface morphology simulation image of at least one of the predicted reaction products, a product simulation image is generated as a training sample for the material analysis model; wherein the image areas in the product simulation image corresponding to each of the surface morphology simulation images all have corresponding label data.

[0006] In some embodiments, for each of the predicted reaction products, generating a surface morphology simulation map of the predicted reaction product based on the molecular structure of the predicted reaction product includes: For each of the predicted reaction products, based on the molecular structure of the predicted reaction product, a spherical model schematic diagram of the predicted reaction product is constructed, and for each of the predicted reaction products, image processing is performed on the spherical model schematic diagram to obtain a surface morphology simulation diagram of the predicted reaction product; or For each of the predicted reaction products, the electron cloud distribution of the predicted reaction product is constructed based on density functional theory, and based on the electron cloud distribution, a surface morphology simulation map of the predicted reaction product is obtained.

[0007] In some embodiments, generating a product simulation map based on the surface morphology simulation map of at least one predicted reaction product comprises: A random number of surface morphology simulation images of different predicted reaction products are added to a preset background image according to the posture parameters of each of the surface morphology simulation images to generate a product simulation image of the selected reactant; wherein the posture parameters are randomly generated.

[0008] In some embodiments, the label data includes pose information, mask information, and key point information of each of the predicted reaction products.

[0009] In some embodiments, determining at least one predicted reaction product of the selected reactant comprises: Obtaining reaction parameters and a reaction mode of the selected reactant; wherein the reaction mode is predetermined based on the chemical properties of the selected reactant; At least one predicted reaction product of the selected reactant is determined based on the reaction parameters and the reaction pattern.

[0010] In some embodiments, determining at least one predicted reaction product of the selected reactant based on the reaction parameters and the reaction pattern comprises: Based on the reaction parameters and the reaction mode, determining a plurality of preliminary reaction products; Eliminate the preliminary reaction products that do not meet the characteristics of organic materials from the multiple preliminary reaction products to obtain the at least one predicted reaction product; and / or for each of the preliminary reaction products, if the preliminary reaction product has a corresponding chiral molecule, use the corresponding chiral molecule as the predicted reaction product.

[0011] In some embodiments, generating a product simulation image based on the surface morphology simulation image of at least one predicted reaction product as a training sample for a material analysis model includes: generating the product simulation map based on the surface morphology simulation map of at least one of the predicted reaction products; Based on the style information of the real experimental image, the style of the product simulation image is converted to obtain the training sample of the material analysis model.

[0012] In some embodiments, after generating a product simulation image based on the surface morphology simulation image of at least one of the predicted reaction products as a training sample for a material analysis model, the method further includes: The target experimental image is input into the material analysis model to obtain the experimental analysis result corresponding to the target experimental image.

[0013] In addition, to achieve the above objectives, the present application also proposes a sample generation device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the sample generation method described above.

[0014] In addition, to achieve the above objectives, the present application also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the sample generation method described above are implemented.

[0015] One or more technical solutions proposed in this application have at least the following technical effects: By determining at least one predicted reaction product of the selected reactant; for each predicted reaction product, based on the molecular structure of the predicted reaction product, generating a surface morphology simulation map of the predicted reaction product; based on at least one surface morphology simulation map of the predicted reaction product, generating a product simulation map as a training sample for the material analysis model, so that sample acquisition can be independent of real data, and there is no need to manually label a large amount of real experimental data, thereby improving the efficiency of sample acquisition and reducing labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A schematic diagram of a flow chart of a sample generation method provided in an embodiment of the present application is shown; Figure 2 A schematic diagram of the molecular structure of a selected reactant provided by an exemplary embodiment of the present application is shown; Figure 3 A schematic diagram showing a preliminary product provided by an exemplary embodiment of the present application is shown; Figure 4 A schematic diagram of predicting reaction products provided by an exemplary embodiment of the present application is shown; Figure 5 A simulated diagram of the surface morphology of a predicted reaction product provided by an exemplary embodiment of the present application is shown; Figure 6 A schematic diagram of a product simulation diagram provided by an exemplary embodiment of the present application is shown; Figure 7 A schematic diagram showing sample data provided by an exemplary embodiment of the present application is shown; Figure 8 A schematic diagram showing a real experimental image provided by an exemplary embodiment of the present application after being processed by a material analysis model; Fig. 9 A schematic diagram of the structure of a sample generation device provided in an embodiment of the present application is shown.

[0019] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0020] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0021] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0022] The main solutions of the embodiments of the present application are: determining at least one predicted reaction product of the selected reactant; predicting that the reaction product is an organic material; for each predicted reaction product, generating a surface morphology simulation map of the predicted reaction product based on the molecular structure of the predicted reaction product; based on at least one surface morphology simulation map of the predicted reaction product, generating a product simulation map as a training sample for a material analysis model; wherein the image areas in the product simulation map corresponding to each surface morphology simulation map all have corresponding label data.

[0023] In the related technology, experimental image analysis of surface materials mainly relies on machine learning models, which usually require a large amount of labeled data for training. The labeled data is generally performed manually by staff, which is not only time-consuming and labor-intensive, but also costly. In many cases, in order to obtain more detailed analysis results, very fine annotation is required, which is very difficult in actual operation.

[0024] In summary, how to obtain model training samples more conveniently and quickly is a problem that needs to be solved urgently.

[0025] Based on this, the present application provides a solution that enables sample acquisition to be independent of real data and eliminates the need for manual labeling of large amounts of real experimental data, thereby improving the efficiency of sample acquisition and reducing labor costs.

[0026] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or a sample generation device capable of realizing the above functions, etc. The following takes the sample generation device as an example to illustrate this embodiment and the following embodiments.

[0027] Reference Figure 1 , Figure 1 The flow chart of a sample generation method provided by an embodiment of the present application is shown. The sample generation method can be applied to a sample generation device, and includes the following steps S110 to S130: Step S110, determining at least one predicted reaction product of the selected reactant.

[0028] The selected reactant refers to an object to be predicted that is pre-specified by a staff member for predicting a corresponding product, and the selected reactant may be one of the materials for preparing the organic material.

[0029] It is understandable that when obtaining model training samples, the purpose of the material analysis model in the subsequent steps should be clarified first. In this application, the material analysis model is used to perform statistical analysis on real experimental data. In real experiments, for organic materials, probe detection methods such as scanning tunneling microscopy (STM), atomic force microscopy (AFM), and scanning electron microscopy (SEM) are usually used to obtain corresponding experimental images. After obtaining the experimental images, manual labeling, identification, statistics and other operations are generally performed, which is difficult, labor-intensive, and inefficient. To solve this problem, this application designs a training material analysis model, so that the material analysis model can be used to quickly and conveniently identify and analyze experimental images, so as to provide staff with analysis and statistical results.

[0030] To this end, the samples acquired in step S110 to step S130 of this embodiment should at least include reactants, products, and statistical analysis results of the products.

[0031] To obtain the "product", in this embodiment, the reaction parameters of the selected reactant can be obtained; based on the reaction parameters and the reaction pattern, at least one predicted reaction product of the selected reactant is determined; wherein the reaction pattern is predicted based on the physical and chemical properties of the selected reactant.

[0032] Among them, reaction parameters refer to a series of constraints used to constrain the direction of the reaction; it is understandable that for larger molecules, there are many possibilities for chemical reactions to occur, and too many reaction products are not conducive to statistical analysis. In order to analyze and count more accurately, this application designs certain constraints on the direction of the reaction. For example, reaction parameters may include reaction steps, allowed reaction sites, the maximum number of reaction site connections, etc. In addition, reaction parameters can be manually input by staff to complete the setting, and this embodiment is not limited here.

[0033] In a feasible embodiment, multiple preliminary reaction products are determined based on reaction parameters and reaction patterns; preliminary reaction products that do not meet the characteristics of organic materials are eliminated from the multiple preliminary reaction products to obtain at least one predicted reaction product; and / or for each preliminary reaction product, if the preliminary reaction product has a corresponding chiral molecule, the corresponding chiral molecule is used as the predicted reaction product.

[0034] Wherein, the reaction mode refers to at least one of the chemical reaction modes that may occur in the aforementioned selected reactants. It is understandable that chemical reactions include a variety of reaction modes, such as substitution reactions, addition reactions, etc. In this embodiment, the possible reaction modes can be automatically identified by analyzing the physicochemical characteristics of the selected reactants, such as reactive groups, spatial structure symmetry, etc. For example, assuming that the selected reactant carries a halogen group, the halogen group generally has a higher reactivity and an elimination reaction may occur.

[0035] The direction of the chemical reaction is limited by the reaction mode and reaction parameters to ensure that the possible products are within the set range of interest and multiple preliminary reaction products are obtained. However, the preliminary reaction products are not perfect, and may contain structures that do not meet the requirements or repeated units, and some chiral molecules may not be taken into account.

[0036] In some embodiments, the reaction mode can be selected by the experience of the staff, or can be predicted by a machine learning model. If predicted by a machine learning model, in some embodiments, the number of possible reaction modes can also be determined according to the confidence level, such as setting a confidence threshold, so as to ensure that the number of subsequent products is within a controllable range, so as to take into account both prediction accuracy and efficiency.

[0037] Therefore, in some embodiments, the units that do not meet the characteristics of organic materials in the preliminary reaction products can be eliminated, and for the units with chiral molecules, their chiral molecules are introduced as predicted reaction products to obtain the final predicted reaction products. Wherein, chiral molecules refer to molecules with certain configurations or conformations that are different from their mirror images and cannot overlap with each other. In some embodiments, the staff can also perform screening and elimination operations from the preliminary reaction products to obtain products of interest, which are not limited in this embodiment.

[0038] In some other possible implementations, a pre-trained reaction product prediction model can also be used to directly output the predicted reaction product corresponding to the selected reactant. Specifically, a large amount of experimental data of organic materials can be used to pre-train a targeted reaction product prediction model, and the selected reactant is input into the reaction product prediction model to directly obtain the output result. For the output result, operations such as screening, elimination, and addition can also be performed based on the above rules, which will not be repeated in this embodiment.

[0039] Step S120 , for each predicted reaction product, based on the molecular structure of the predicted reaction product, a surface morphology simulation image of the predicted reaction product is generated.

[0040] The surface morphology simulation image refers to a simulated image generated for predicting reaction products by imitating the style of a single molecule in the experimental image taken from the aforementioned real experiment.

[0041] In a feasible implementation, for each predicted reaction product, a spherical model schematic diagram of the predicted reaction product can be constructed based on the molecular structure of the predicted reaction product; for each predicted reaction product, the spherical model schematic diagram is image processed to obtain a surface morphology simulation diagram of the predicted reaction product.

[0042] Among them, the sphere model is a model commonly used in chemistry to visualize molecular structures. Each atom in a molecule is represented by a sphere, and the size of different spheres is proportional to the diameter of the corresponding atom. By establishing a schematic diagram of the sphere model for predicting reaction products, it is convenient to generate simulation images that are closer to real experimental images.

[0043] After obtaining the schematic diagram of the ball model, in some implementations, the schematic diagram of the ball model can also be processed by a certain image processing algorithm, such as blurring boundaries, adding noise points, etc., to obtain the final surface morphology simulation image, thereby achieving a further simulation effect.

[0044] In other feasible implementations, simulations can also be performed based on specific chemical theories, such as density functional theory (DFT), to construct an electron cloud distribution for predicting reaction products, and then based on the electron cloud distribution, a surface morphology simulation map that conforms to physical a priori can be further obtained. This process can be manually modeled by staff, or it can be automatically modeled and generated by the sample generation device provided in this application, and this embodiment is not limited here.

[0045] Step S130, based on at least one surface morphology simulation image of the predicted reaction product, a product simulation image is generated as a training sample for the material analysis model.

[0046] The product simulation image refers to a simulated image of the current predicted reaction product as a whole, which is generated by imitating the experimental image taken by the above-mentioned real experiment. The product simulation image includes at least one predicted reaction product, and the number of each predicted reaction product can be different.

[0047] As mentioned above, the material analysis model is used to analyze and count the products in the experimental image. Therefore, when acquiring samples, this embodiment is designed to label each product in the product simulation image with a corresponding real label, so as to facilitate the learning of the material analysis model. Specifically, the image areas corresponding to each surface morphology simulation image in the product simulation image have corresponding label data, and the label data may include but is not limited to posture information, mask information, and key point information.

[0048] In some embodiments, the pose information is used to characterize the position and angle of the corresponding image area; the mask information is used to characterize the mask of the corresponding image area; and the key point information can be used to characterize the boundary of the corresponding image area.

[0049] In some embodiments, in order to ensure the authenticity and versatility of the subsequent statistical analysis results, all the contents in the label data can be randomly processed. For example, in this embodiment, a random number of surface morphology simulation images of different predicted reaction products can be added to a preset background image according to the posture parameters of each surface morphology simulation image to generate a product simulation image of the selected reactant; wherein the posture parameters are randomly generated.

[0050] After obtaining the surface morphology simulation image, the surface morphology simulation image can be further converted to make it closer to the actual experimental situation. Specifically, based on the style information of the real experimental image, the product simulation image can be converted to a style, and image processing operations such as adjusting brightness and contrast, adding noise, etc. can be performed to obtain training samples that are closer to the real experimental image.

[0051] In other feasible implementations, the style transfer model can be pre-trained based on a large amount of prior data, which will not be described in detail here. After obtaining a product simulation image formed by combining multiple surface morphology simulation images corresponding to the predicted reaction products, the product simulation image can be input into the style transfer model to directly obtain the final training sample.

[0052] In some embodiments, after obtaining the training samples, the material analysis model can be trained using the training samples. In some embodiments, to further improve the robustness of the material analysis model, a detection head can be trained by adding some non-target reaction products, thereby increasing the model's ability to identify non-target products.

[0053] Specifically, the predicted reaction products obtained by simulation prediction may not cover all product categories in reality, such as impurities and other possible products that are not included in the predicted reaction products. These products that are not taken into account (out-of-domain products) may cause the analysis and recognition performance of the model to deteriorate. In order to increase the detection capability of these out-of-domain products and enhance the performance of the model within the domain, at least one detection head can be added to the material analysis model. The detection head can output a prediction value based on the label information. For any product, if it belongs to the label within the domain, the prediction value is 1; if it does not belong to the label within the domain, the prediction value is 0. The detection head can be trained by generating some non-target reaction products or using some irregular patterns. Finally, the material analysis model can not only output product categories, but also provide information on non-domain targets.

[0054] After obtaining the material analysis model, the material analysis model can be applied to specific experimental data analysis, and the specific examples can be referred to as follows. In some embodiments, during or after the actual experiment, the acquired target experimental image can be input into the material analysis model to obtain the experimental analysis result corresponding to the target experimental image.

[0055] As mentioned earlier, through training, the material analysis model can automatically analyze and identify the products in the image, and statistically classify the products under the same label, so as to obtain some analytical statistical results, such as the overall product distribution in the current experimental environment, the property distribution of the products and other key information, so as to facilitate the staff to make decisions for the experiment or analyze the experimental data.

[0056] Based on this, staff no longer need to perform manual identification and manual labeling of experimental images, and can achieve more convenient analysis and statistics, greatly improving work efficiency and saving labor costs.

[0057] The present embodiment provides a sample generation method, by determining at least one predicted reaction product of a selected reactant; for each predicted reaction product, generating a surface morphology simulation image of the predicted reaction product based on the molecular structure of the predicted reaction product; based on at least one of the surface morphology simulation images of the predicted reaction product, generating a product simulation image as a training sample for a material analysis model, thereby enabling sample acquisition to be independent of real data and eliminating the need for manual annotation of a large amount of real experimental data, thereby improving the efficiency of sample acquisition and reducing labor costs.

[0058] For example, in order to help understand the implementation process of the sample generation method obtained by combining this embodiment with the above embodiments, it is assumed that the selected reactant is Figure 2 The molecular structure in (hereinafter referred to as the target molecule), specifically: A. Product prediction (1) Determine the reaction mode: Based on the molecular structure and chemical properties of the target molecule, it is known that the carbon atom (C) next to the bromine atom (Br) and the nitrogen atom (N) in the BP molecule has a higher reactivity. Therefore, the following two reaction modes are given priority: bromine-bromine (Br-Br) reaction between two molecules, that is, the bromine atoms are removed separately and the corresponding carbon atoms form bonds; one molecule removes the bromine atom, and the corresponding carbon atom forms a bond with the carbon atom next to the nitrogen atom of another molecule.

[0059] (2) Determine the reaction parameters: Based on the above reaction mode, other possible connections are not considered and the maximum number of reaction steps is 3.

[0060] (3) Based on the above (1) and (2), the predicted reaction product is obtained: ① The initial reaction products are as follows Figure 3 As shown, the first row contains two reaction molecules; the second row contains 3, and the third row contains 4. The specific method is that the molecules in the first row react through Br-Br; the second row is the product of the first row and the reaction molecule through Br-C reaction; the third row is the product of the second row and the reaction molecule through Br-C reaction.

[0061] ② Eliminate repeated units and units that do not meet the requirements of organic materials, and add chiral molecules, such as Figure 4 As shown, the final predicted reaction product was obtained.

[0062] B. Generate training samples (1) Using the spherical model, the surface morphology simulation diagrams of each predicted reaction product are established, such as Figure 5 shown.

[0063] (2) Generate a complete product simulation graph, which contains the labels of each individual unit, such as Figure 6 shown.

[0064] (3) Perform style conversion on the complete product simulation image to make it closer to the real experimental style, such as Figure 7 shown.

[0065] C. Training the material analysis model Specifically, the training is based on the yolov11-obb model. The training method is consistent with the conventional target detection model training, and the training is performed by providing pictures and corresponding label files as samples. Among them, for the rotation detection frame, the label file is the coordinates of the four corner points [x1, y1, x2, y2, x3, y3, x4, y4].

[0066] Through the material analysis model, based on the experimental images, we can determine which products appear in a high proportion among the predicted products supported by the model, as well as the corresponding angles of the products.

[0067] D. Use the trained material analysis model to analyze real experimental images.

[0068] After inputting the real experimental image into the material analysis model for analysis, we get Figure 8 The image shown. By analyzing and counting it, we can know the number of occurrences of different products, deflection angles and other information, thus realizing the analysis and statistics of experimental data.

[0069] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the sample generation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0070] The present application provides a sample generation device, which includes: at least one processor; and a memory that is communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the sample generation method in the above-mentioned embodiment 1.

[0071] Reference below Fig. 9 , which shows a schematic diagram of the structure of a sample generation device suitable for implementing the embodiment of the present application. The sample generation device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig. 9The sample generation device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0072] like Fig. 9 As shown, the sample generation device 100 may include a processing device 110 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 120 or a program loaded from a storage device 130 to a random access memory (RAM: Random Access Memory) 140. In RAM140, various programs and data required for the operation of the sample generation device are also stored. The processing device 110, ROM120 and RAM140 are connected to each other through a bus 150. An input / output (I / O) interface 160 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 160: an input device 170 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 180 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 130 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 190. The communication device 190 can allow the sample generation device to communicate with other devices wirelessly or by wire to exchange data. Although the sample generation device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0073] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device, or installed from the storage device 130, or installed from the ROM 120. When the computer program is executed by the processing device 110, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0074] The sample generation device provided by the present application adopts the sample generation method in the above embodiment, which can solve the technical problem of how to obtain model training samples more conveniently and quickly. Compared with the prior art, the beneficial effects of the sample generation device provided by the present application are the same as the beneficial effects of the sample generation method provided by the above embodiment, and the other technical features in the sample generation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0075] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0076] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0077] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the sample generation method in the above-mentioned embodiment.

[0078] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0079] The computer-readable storage medium may be included in the sample generating device; or may exist independently without being assembled into the sample generating device.

[0080] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the sample generation device, the sample generation device can write computer program codes for performing the operations of the present application in one or more programming languages ​​or a combination thereof. The programming languages ​​include object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0081] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0082] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0083] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned sample generation method, and can solve the technical problem of how to obtain model training samples more conveniently and quickly. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the sample generation method provided in the above-mentioned embodiment, and will not be repeated here.

[0084] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A sample generation method, characterized in that: The sample generation method comprises: determining at least one predicted reaction product of the selected reactant; the predicted reaction product being an organic material; For each of the predicted reaction products, generating a surface morphology simulation image of the predicted reaction product based on the molecular structure of the predicted reaction product; Based on the surface morphology simulation image of at least one of the predicted reaction products, a product simulation image is generated as a training sample for the material analysis model; wherein the image areas in the product simulation image corresponding to each of the surface morphology simulation images all have corresponding label data.

2. The sample generation method according to claim 1, characterized in that: For each of the predicted reaction products, generating a surface morphology simulation map of the predicted reaction product based on the molecular structure of the predicted reaction product includes: For each of the predicted reaction products, based on the molecular structure of the predicted reaction product, a spherical model schematic diagram of the predicted reaction product is constructed, and for each of the predicted reaction products, image processing is performed on the spherical model schematic diagram to obtain a surface morphology simulation diagram of the predicted reaction product; or For each of the predicted reaction products, the electron cloud distribution of the predicted reaction product is constructed based on density functional theory, and based on the electron cloud distribution, a surface morphology simulation map of the predicted reaction product is obtained.

3. The sample generation method according to claim 1, characterized in that: The step of generating a product simulation map based on the surface morphology simulation map of at least one predicted reaction product comprises: A random number of surface morphology simulation images of different predicted reaction products are added to a preset background image according to the posture parameters of each of the surface morphology simulation images to generate a product simulation image of the selected reactant; wherein the posture parameters are randomly generated.

4. The sample generation method according to claim 1, characterized in that: The label data includes pose information, mask information and key point information of each predicted reaction product.

5. The sample generation method according to claim 1, characterized in that: Determining at least one predicted reaction product of the selected reactant comprises: Obtaining reaction parameters and a reaction mode of the selected reactant; wherein the reaction mode is predetermined based on the chemical properties of the selected reactant; At least one predicted reaction product of the selected reactant is determined based on the reaction parameters and the reaction pattern.

6. The sample generation method according to claim 5, characterized in that: The step of determining at least one predicted reaction product of the selected reactant based on the reaction parameters and the reaction pattern comprises: Based on the reaction parameters and the reaction mode, determining a plurality of preliminary reaction products; Eliminate the preliminary reaction products that do not meet the characteristics of organic materials from the multiple preliminary reaction products to obtain the at least one predicted reaction product; and / or for each of the preliminary reaction products, if the preliminary reaction product has a corresponding chiral molecule, use the corresponding chiral molecule as the predicted reaction product.

7. The sample generation method according to claim 1, characterized in that: The step of generating a product simulation image based on the surface morphology simulation image of at least one predicted reaction product as a training sample for a material analysis model comprises: generating the product simulation map based on the surface morphology simulation map of at least one of the predicted reaction products; Based on the style information of the real experimental image, the style of the product simulation image is converted to obtain the training sample of the material analysis model.

8. The sample generation method according to claim 1, characterized in that: After generating a product simulation image based on the surface morphology simulation image of at least one predicted reaction product as a training sample for a material analysis model, the method further includes: The target experimental image is input into the material analysis model to obtain the experimental analysis result corresponding to the target experimental image.

9. A sample generating device, characterized in that: The sample generation device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the sample generation method according to any one of claims 1 to 8.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the sample generation method according to any one of claims 1 to 8 are implemented.

Images