Algae microscopic image automatic labeling method and device, computer and storage medium

Abstract

The invention discloses an algae microscopic image automatic labeling method and device, a computer and a storage medium, all manual labeling data and automatic labeling data are merged to form a labeling database, image materials in the labeling database are cut and / or spliced to form a comprehensive training set, and a labeling model is updated according to the comprehensive training set. According to the algae microscopic image automatic labeling method and device, the computer and the storage medium provided by the invention, a plurality of individual pure algae samples of different algae are rapidly obtained by expanding culture of the sample water sample, so that a labeling model of specific algae can be rapidly trained by utilizing light-weight single algae manual labeling data, and the single algae microscopic image is automatically labeled; according to the method, large-scale single algae labeling data can be quickly and accurately obtained by utilizing an automatic labeling function of the labeling model, a mixed algae sample is simulated by cutting and splicing different algae images, a training set can be efficiently formed, and the labeling model capable of identifying a complex algae sample target can be quickly trained.

Description

Technical Field

[0001] The present invention relates to the field of algae annotation, and particularly to an automatic annotation method, device, computer, and storage medium for algae microscopic images. Background Art

[0002] The types and quantities of algae are important indicators of the water ecosystem, and algae microscopic image analysis plays an important role in algae population research, water quality monitoring, and environmental assessment. However, the existing algae annotation datasets mainly rely on manual annotation, which is time-consuming and highly subjective, restricting the rapid construction of high-quality datasets.

[0003] Deep learning object detection algorithms (such as YOLO, etc.) have made remarkable progress in object detection tasks. These algorithms have shown great potential in constructing algae object detection datasets. However, due to the diversity of algae species, the microscopic image detection of multi-algae objects faces huge challenges. Existing object detection technologies have certain difficulties in accurately classifying different algae objects and providing effective annotation information. In the field of algae annotation, there is still a lack of an efficient automatic annotation method, making the rapid acquisition of training data a bottleneck restricting the deep learning research of algae identification. Summary of the Invention

[0004] The main purpose of the present invention is to provide an automatic annotation method, device, computer, and storage medium for algae microscopic images, aiming to solve the problem that in the field of algae annotation, there is still a lack of an efficient automatic annotation method, making the rapid acquisition of training data a bottleneck restricting the deep learning research of algae identification.

[0005] To achieve the above purpose, the present invention provides an automatic annotation method for algae microscopic images, including: S1. Receiving a microscopic image of a sub-water sample, wherein the sub-water sample is formed by diluting a sample water sample, splitting it, and culturing it, and the microscopic image corresponds to the sub-water sample; S2. Dividing all the microscopic images into artificial group image data and automatic group image data; S3. Receiving artificial annotation data and incorporating it into the training set, and updating the annotation model through the training set, wherein the artificial annotation data is the annotation formed by an artificial person for all the microscopic images in the artificial group image data; S4. Annotating the automatic group image data according to the annotation model to form automatic annotation data; S5. Repeating the above steps S1 to S4 several times, and merging all the artificial annotation data and the automatic annotation data to form an annotation database, wherein the annotation database includes image materials and material annotation attributes corresponding to the image materials; S6. Crop and / or splice the image materials in the annotation database to form a comprehensive training set, and update the annotation model according to the comprehensive training set.

[0006] Further, after the step of S1, it includes: Receive the merged microscopic images of the merged water samples, where the merged water samples are formed by combining all the sub-water samples in pairs, and the merged microscopic images correspond to the merged water samples; The step of S6 further includes: Obtain fused annotation data according to the merged microscopic images and the material annotation attributes corresponding to the merged microscopic images; Train and update the annotation model according to the fused annotation data.

[0007] Further, in the step of S6, determine whether to crop or splice the image materials according to the image materials and the material annotation attributes corresponding to the image materials.

[0008] Further, in the step of S5, give the recommended number of split copies corresponding to the sample water sample according to the manual annotation data in the previous round of steps S1 to S4.

[0009] Further, in the step of S5, give the recommended cultivation time corresponding to the sub-water sample according to the manual annotation data in the previous round of steps S1 to S4.

[0010] Further, in the step of S2, the quantity ratio between the manual group image data and the automatic group image data is between 1:2 and 1:10.

[0011] Further, in the step of S5, the number of repetitions of steps S1 to S4 is 2 to 5 times.

[0012] The present invention also provides a device for running the above method, including: A storage module, configured to receive the microscopic images of the sub-water samples, where the sub-water samples are formed after diluting the sample water sample, splitting it, and culturing it, and the microscopic images correspond to the sub-water samples; A classification module, configured to divide all the microscopic images into manual group image data and automatic group image data; A first training and forming module, configured to receive the manual annotation data and incorporate it into the training set, and update the annotation model through the training set, where the manual annotation data is the annotation formed by manual operation on all the microscopic images in the manual group image data; A first training working module, configured to perform annotation on the automatic group image data according to the annotation model to form automatic annotation data; A library formation module, configured to repeat the steps of S1 to S4 for more than a certain number of times, and merge all the manually labeled data and the automatically labeled data to form a labeled database, where the labeled database includes image materials and material labeling attributes corresponding to the image materials; A second training formation module, configured to crop and / or splice the image materials in the labeled database to form a comprehensive training set, and update the labeling model according to the comprehensive training set.

[0013] The present invention also provides a computer, including a memory and a processor, where a computer program is stored in the memory, and when the processor executes the computer program, the steps of the above-mentioned automatic labeling method for algal microscopic images are implemented.

[0014] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned automatic labeling method for algal microscopic images are implemented.

[0015] The automatic labeling method, device, computer, and storage medium for algal microscopic images provided by the present invention can quickly obtain multiple pure algal samples of different algae through sample water sample expansion culture, so that a labeling model for specific algae can be quickly trained by using lightweight single-algae manually labeled data to automatically label single-algae microscopic images; by using the automatic labeling function of the labeling model, large-scale single-algae labeled data can be quickly and accurately obtained, and a training set can be efficiently formed by cropping and splicing different algae images to simulate mixed algal samples, and a labeling model capable of identifying complex algal sample targets can be quickly trained; it overcomes the defects of the automatic labeling model in classification work. In the background of single algae, only a small amount of manually labeled data set needs to be provided to efficiently form a labeling model, and then large-scale automatic labeling work can be realized, which can not only promote the development of intelligent algal identification technology, but also has important significance for the research and application in related fields. Description of the Drawings

[0016] Figure 1 is a schematic diagram of the steps of the automatic labeling method for algal microscopic images according to an embodiment of the present invention; Figure 2 is a conceptual schematic diagram of the device for running the automatic labeling method for algal microscopic images according to an embodiment of the present invention.

[0017] The realization, functional characteristics, and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] Those skilled in the art of the present technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the", "above-mentioned" and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the specification of the present invention means the presence of the described features, integers, steps, operations, elements, units, modules and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, units, modules, components and / or their groups. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0020] Those skilled in the art of the present technology can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention pertains. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0021] Referring to Figure 1 , in an embodiment of the present invention, a method for automatically annotating microscopic images of algae includes: S1. Receiving microscopic captured images of sub-water samples, where the sub-water samples are formed by diluting sample water samples, splitting them, and culturing them, and the microscopic captured images correspond to the sub-water samples; S2. Dividing all the microscopic captured images into artificial group image data and automatic group image data; S3. Receiving artificial annotation data and incorporating it into the training set, and updating the annotation model through the training set, where the artificial annotation data is the annotation formed by humans for all the microscopic captured images in the artificial group image data; S4. Annotating the automatic group image data according to the annotation model to form automatic annotation data; S5. Repeating the above steps S1 to S4 several times, and merging all the artificial annotation data and the automatic annotation data to form an annotation database, where the annotation database includes image materials and material annotation attributes corresponding to the image materials; S6. Cropping and / or splicing the image materials in the annotation database to form a comprehensive training set, and updating the annotation model according to the comprehensive training set.

[0022] In the prior art, in the field of algae annotation, there is still a lack of an efficient automatic annotation method, making the rapid acquisition of training data a bottleneck restricting the deep learning research of algae identification.

[0023] In the present invention, in step S1, a microscopic image of a sub-water sample is received, where the sub-water sample is formed after diluting a sample water sample, splitting it, and culturing it. The microscopic image corresponds to the sub-water sample. In a typical process, the diluted sample water sample is transferred to a 96-well plate for amplification culture. The dilution degree ensures that there is at most one algae individual in each well. After a period of cultivation, multiple individuals of a single algae can be obtained in each well.

[0024] In step S2, all microscopic images are divided into artificial group image data and automatic group image data. The artificial group image data is manually annotated in the subsequent process, and the automatic group image data is automatically annotated through an annotation model in the subsequent process. The ratio range between the artificial group image data and the automatic group image data can be relatively large, such as between 1:1 and 1:20.

[0025] In step S3, the manually annotated data is received and incorporated into the training set, and the annotation model is updated through the training set. The manually annotated data is the annotation formed by manually annotating all microscopic images in the artificial group image data. When there is no annotation model yet, the process of updating the annotation model refers to generating the annotation model.

[0026] In step S4, the automatic group image data is annotated according to the annotation model to form automatic annotation data. In this process, the annotation of the automatic group image data can be quickly completed through the annotation model.

[0027] In steps S5 and S6, the above steps S1 to S4 are repeated several times, and all the manually annotated data and the automatic annotation data are combined to form an annotation database. The annotation database includes image materials and material annotation attributes corresponding to the image materials. The image materials in the annotation database are trimmed and / or spliced to form a comprehensive training set, and the annotation model is updated according to the comprehensive training set. Through the formation of the comprehensive training set, a relatively complex training process can be carried out on the annotation model, and the above training process does not require the introduction of manual labor. In the selection process of trimming and splicing the image materials, certain logical rules can be set. For example, if the material annotation attribute of the image material shows only one type of algae, then the image material can be trimmed and spliced. If the material annotation attribute of the image material shows multiple types of algae, then the image material can be spliced. It should be particularly noted that during the process of repeating steps S1 to S4, the ratio of dividing the microscopic images into artificial group image data and automatic group image data is not limited to a fixed value. Even all the microscopic images can be divided into automatic group image data, thereby avoiding the consumption of manual labor. At this time, the annotation model cannot be updated in step S3.

[0028] In one implementation step: (1) Collect field water samples, dilute them with an algae culture medium, and transfer them to a 96-well plate for expansion culture. The dilution degree ensures that there is at most one algae individual in each well. After a period of cultivation, multiple individuals of a single algae can be obtained in each well.

[0029] (2) Collect single algae from different wells and take microscopic images under a microscope. The microscopic images are divided into artificial group image data and automatic group image data in a ratio of 1:5. Since there is only single algae, the artificial annotation efficiency is higher than that of complex field samples.

[0030] (3) For single algae, use the artificial annotation data for the artificial group image data as the training set, combined with the yolov11 model. Since there is only single algae, an effective annotation model for this algae can be quickly established, and then accurate and rapid automatic annotation work can be performed.

[0031] Repeat the above process from (1) to (3), replace the artificial annotation with the automatic annotation of the optimized annotation model, quickly obtain a large number of annotated images and annotation data, form mixed algae images through the cropping and splicing of different category algae images, generate a large-scale training set, and provide a basis for the target detection of further complex algae samples.

[0032] In summary, by expanding the culture of the sample water sample, multiple pure algae samples of different algae can be quickly obtained, so that an annotation model for a specific algae can be quickly trained by using lightweight single algae manual annotation data, and automatic annotation can be performed on the microscopic images of single algae; by using the automatic annotation function of the annotation model, large-scale single algae annotation data can be quickly and accurately obtained, and mixed algae samples can be simulated through the cropping and splicing of different algae images, and a training set can be efficiently formed, and an annotation model capable of identifying the targets of complex algae samples can be quickly trained; the defects of the automatic annotation model in the classification work are overcome. In the background of single algae, only a small amount of artificial annotation data set needs to be provided to efficiently form an annotation model, and then large-scale automatic annotation work can be realized, which can not only promote the development of intelligent algae recognition technology, but also has important significance for the research and application of related fields.

[0033] In one embodiment, after the step of S1, it includes: Receiving the combined microscopic images of the combined water samples, where the combined water samples are formed by pairwise combining all the sub-water samples, and the combined microscopic images correspond to the combined water samples; The step of S6 further includes: Obtaining fused annotation data according to the combined microscopic images and the material annotation attributes corresponding to the combined microscopic images; Train and update the annotation model according to the fused annotation data.

[0034] In this embodiment, after obtaining the microscopic images in step S1, all sub-water samples can be combined in pairs to provide a basis for subsequent work, and the combined microscopic images are obtained. Instead of performing automatic annotation on the combined microscopic images, the annotation results (material annotation attributes) of the two sub-water samples that make up the combined water sample are used to merge as the annotation result (fused annotation data) of the combined water sample. According to the fused annotation data, the annotation model is trained and updated. The above methods of combining water samples and fused annotation data can, on the premise that the workload hardly increases, increase the training library capacity while doubling the complexity of the provided fused annotation data, providing a basis for the training and update of the annotation model.

[0035] In one embodiment, in step S6, determine whether to crop or splice the image material according to the image material and the material annotation attribute corresponding to the image material.

[0036] In step S6, a comprehensive training set needs to be formed, and the way to form the comprehensive training set can be cropping or splicing, but no evaluation criteria for choosing which method are given, so the formation of the comprehensive training set is not optimized. In this embodiment, considering that the annotation database contains all the image materials and the material annotation attributes corresponding to the image materials, for example, the material annotation attribute corresponding to a certain image material shows that there are 3 different algae, and their respective quantities are 8, 9, and 7. Then the recommended processing method for this image material is cropping. For example, the material annotation attribute corresponding to a certain image material shows that there is 1 type of algae and its quantity is 2. Then the recommended processing method for this image material is splicing. Specifically, determine whether to crop or splice the image material according to the types and quantities of algae in the image material and the material annotation attribute corresponding to the image material.

[0037] In one embodiment, in step S5, give the recommended number of split copies of the sample water sample according to the manual annotation data in the previous round of steps S1 to S4.

[0038] In this embodiment, considering that steps S1 to S4 need to be repeated multiple times, in order to satisfy multiple individuals with a single type of algae in the sub-water samples (the efficiency and accuracy of manual annotation and automatic annotation are relatively high), the number of split copies of the sample water sample in the next round of steps S1 to S4 can be given according to the situation of the manual annotation data. For example, in the previous round of steps S1 to S4, if the manual annotation data shows that the types of algae are too many, then give a recommended reduced number of split copies according to the specific situation; for example, in the previous round of steps S1 to S4, if the manual annotation data shows that the types of algae are too few, then give a recommended increased number of split copies according to the specific situation.

[0039] In one embodiment, in the step S5, a recommended incubation time corresponding to the sub-water sample is given based on the manually labeled data in the previous round of steps S1 to S4.

[0040] In this embodiment, considering that steps S1 to S4 need to be repeated multiple times, in order to meet the requirement that the components in the sub-water sample are multiple individuals of a single algae (both manual and automatic annotation have high efficiency and accuracy), the recommended incubation time for the sub-water sample in the next round of steps S1 to S4 can be given according to the situation of the manually labeled data. For example, in the previous round of steps S1 to S4, if the manually labeled data shows that the number of algae is too large, a reduced recommended incubation time is given according to the specific situation; for example, in the previous round of steps S1 to S4, if the manually labeled data shows that the number of algae is too small, an increased recommended incubation time is given according to the specific situation.

[0041] In one embodiment, in step S2, the ratio of the number of the manually grouped image data to the number of the automatically grouped image data is between 1:2 and 1:10.

[0042] In this embodiment, a suitable ratio of manually grouped image data to automatically grouped image data is given, so as to improve the efficiency and accuracy of the overall annotation without high manual workload.

[0043] In one embodiment, in step S5, steps S1 to S4 are repeated 2 to 5 times.

[0044] In this embodiment, a suitable number of repetitions of S1 to S4 is given to improve the efficiency and accuracy of the overall labeling without high manual workload.

[0045] The present invention also provides a device for executing the above method, comprising: The storage module 10 is used to receive a microscopic image of a sub-water sample, wherein the sub-water sample is formed by diluting the sample water sample and then splitting and culturing it, and the microscopic image corresponds to the sub-water sample; A classification module 20, used for classifying all the microscopic images into manually grouped image data and automatically grouped image data; A first training forming module 30 is used to receive manually annotated data and incorporate it into a training set, and to update the annotation model through the training set, wherein the manually annotated data is manually annotated on all the microscopic images in the manual group image data; A first training working module 40, used for annotating the automatic group image data according to the annotation model to form automatic annotation data; A library formation module 50, configured to repeat the steps of S1 to S4 for more than a certain number of times, and merge all the manually labeled data and the automatically labeled data to form a labeled database, wherein the labeled database includes image materials and material labeling attributes corresponding to the image materials; A second training formation module 60, configured to crop and / or splice the image materials in the labeled database to form a comprehensive training set, and update the labeling model according to the comprehensive training set.

[0046] In this embodiment, the working mode of the device will not be elaborated here. Refer to the foregoing method content.

[0047] The present invention also provides a computer, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the above-mentioned automatic labeling method for algal microscopic images are implemented.

[0048] The computer device includes a processor, a storage, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The storage of the computer device includes a non-volatile storage medium and an internal storage. The non-volatile storage medium stores an operating system and a computer program. The internal storage provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, a file storage method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0049] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned automatic labeling method for algal microscopic images are implemented.

[0050] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in this application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, there are various forms of RAM, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0051] In summary, the automatic annotation method, device, computer, and storage medium for algal microscopic images provided by the present invention can quickly obtain multiple pure algal samples of different algae through the expansion culture of the sample water sample, so as to quickly train an annotation model for a specific alga by using lightweight single-alga manually annotated data and automatically annotate the single-alga microscopic images; utilize the automatic annotation function of the annotation model to quickly and accurately obtain a large amount of single-alga annotation data, and simulate a mixed algal sample by cropping and splicing different algal images, so as to efficiently form a training set and quickly train an annotation model capable of identifying the targets of complex algal samples; overcome the defects of the automatic annotation model in the classification work. In the background of a single alga, only a small amount of manually annotated data set needs to be provided to efficiently form an annotation model, and then large-scale automatic annotation work can be realized, which can not only promote the development of intelligent algal recognition technology, but also has important significance for the research and application in related fields.

[0052] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be similarly included in the patent protection scope of the present invention.

Claims

1. An automatic annotation method for microscopic images of algae, characterized in that, include: S1, receiving a microscopic image of a sub-water sample, wherein the sub-water sample is formed by diluting a sample water sample and then splitting and culturing it, and the microscopic image corresponds to the sub-water sample; S2, dividing all the microscopic images into manual group image data and automatic group image data; S3, receiving manually annotated data and incorporating them into a training set, and updating the annotation model through the training set, wherein the manually annotated data is manually annotated on all the microscopic images in the manual group of image data; S4, annotating the automatic group image data according to the annotation model to form automatic annotation data; S5, repeating the above steps S1 to S4 several times, and merging all the manual annotation data and the automatic annotation data to form an annotation database, wherein the annotation database includes image materials and material annotation attributes corresponding to the image materials; S6. Cutting and / or splicing the image materials in the annotation database to form a comprehensive training set, and updating the annotation model according to the comprehensive training set.

2. The automatic annotation method for microscopic images of algae according to claim 1, characterized in that, The step of S1 then includes: receiving a combined microscopic image of a combined water sample, wherein the combined water sample is formed by combining all the sub-water samples in pairs, and the combined microscopic image corresponds to the combined water sample; The step of S6 further includes: Obtaining fused annotation data according to the combined microscopic image and the material annotation attribute corresponding to the combined microscopic image; The annotation model is trained and updated according to the fused annotation data.

3. The automatic annotation method for microscopic images of algae according to claim 1, characterized in that In the step S6, it is determined whether the image material is cropped or spliced ​​according to the image material and the material annotation attribute corresponding to the image material.

4. The automatic annotation method for microscopic images of algae according to any one of claims 1 to 3, characterized in that, In the step S5, a recommended number of splits corresponding to the sample water sample is given based on the manually labeled data in the previous round of steps S1 to S4.

5. The automatic annotation method for microscopic images of algae according to any one of claims 1 to 3, characterized in that, In the step S5, a suggested incubation time corresponding to the sub-water sample is given based on the manually labeled data in the previous round of steps S1 to S4.

6. The automatic annotation method for microscopic images of algae according to any one of claims 1 to 3, characterized in that, In the step S2, the ratio of the number of the manually grouped image data to the number of the automatically grouped image data is between 1:2 and 1:

10.

7. The automatic annotation method for microscopic images of algae according to any one of claims 1 to 3, characterized in that, In the step S5, the steps S1 to S4 are repeated 2 to 5 times.

8. An apparatus for operating the method according to any one of claims 1 to 7, characterized in that, include: A storage module (10) is used to receive a microscopic image of a sub-water sample, wherein the sub-water sample is formed by diluting the sample water sample and then splitting and culturing it, and the microscopic image corresponds to the sub-water sample; A classification module (20), used for classifying all the microscopic images into manually grouped image data and automatically grouped image data; A first training forming module (30) is used to receive manually annotated data and incorporate it into a training set, and to update the annotation model through the training set, wherein the manually annotated data is manually annotated on all the microscopic images in the manual group image data; A first training working module (40) is used to annotate the automatic group image data according to the annotation model to form automatic annotation data; A library formation module (50) is configured to repeat the steps of S1 to S4 for more than several times, and merge all the manually annotated data and the automatically annotated data to form an annotation database, wherein the annotation database includes image materials and material annotation attributes corresponding to the image materials; A second training formation module (60) is configured to crop and / or splice the image materials in the annotation database to form a comprehensive training set, and update the annotation model according to the comprehensive training set.

9. A computer, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that, When the processor executes the computer program, the steps of the automatic annotation method for algal microscopic images according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the automatic annotation method for algal microscopic images according to any one of claims 1 to 7 are implemented.

Images