AI training and detection method for identifying cell-derived species
By using AI training and detection methods for different cell types, the problem of automatic species identification has been solved, achieving efficient and accurate species identification and cell content calculation, which is applicable to food safety and scientific research fields.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENZHEN ANLV MEDICAL TECH CO LTD
- Filing Date
- 2024-11-08
- Publication Date
- 2026-05-08
AI Technical Summary
Existing technologies struggle to automatically identify different types of species, especially in food safety and scientific research where species identification is inefficient and costly, and traditional testing methods are time-consuming.
By training AI on labeled images of cell types from two different species, an AI dataset of cell features for different species is obtained. Combined with corresponding AI algorithm modules, the identification of cells from different species can be achieved.
It improves the efficiency and accuracy of species identification, enables rapid identification of cells from different species, reduces detection costs, and provides efficient scientific research and reconnaissance support.
Smart Images
Figure CN121999482A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of formed element analysis technology based on magnified microscopic images, and specifically relates to an AI training and detection method for identifying cell-derived species, as well as a computing, processing, and storage device. Background Technology
[0002] Blood cell characteristics vary among different species. For example, red blood cells are the most numerous type of blood cell. In most vertebrates, mammalian red blood cells are disc-shaped with a concave center; human red blood cells are also disc-shaped with convex edges and a concave center. Abnormalities in red blood cells are often indicators of diseases. However, for species at different evolutionary levels, the morphology and characteristics of their red blood cells differ. For instance, most non-mammalian red blood cells are nucleated.
[0003] The applicant has filed a series of Chinese patents, such as
[0004] 1. CN2022104799126, "A method for rapid focusing and a method for detecting microscopic images in a microscopic image acquisition device";
[0005] 2. CN2023100423151, "Sample Imaging Analysis System and Method";
[0006] 3. CN2023116834572, "AI training method and computing and storage device for white blood cell recognition in samples"; a novel AI recognition technology is used to measure the content of formed elements in samples.
[0007] However, how to identify different types of species based on the above technologies is a technical problem to be solved.
[0008] In existing technologies, the detection is based on known animal types, and the animal type needs to be input before analysis. How to enable the device to automatically identify the species origin of the sample is a technical problem to be solved.
[0009] In the field of food safety, for example, routine or random testing of the species of food sold needs to be conducted, but traditional testing technologies are difficult to implement and require long testing times.
[0010] In scientific research or reconnaissance, many scenarios require the identification and characterization of species based on small blood samples. There is currently no efficient way to do this, as DNA testing is time-consuming, costly, and inefficient. Summary of the Invention
[0011] In this application, the inventors propose to use labeled images of cell types from two different species for AI training to obtain an AI species cell feature dataset. The AI species cell feature dataset, combined with the corresponding AI algorithm module, can identify and distinguish cells of species A from cells of species B. By using advanced AI computing power to complete the identification of different species, species identification based on cell features becomes possible, and species identification becomes more efficient and accurate.
[0012] The technical solution of this application to solve the above-mentioned technical problems is an AI training method for identifying cell-derived species. Microscopic sample image A includes images of cell type A of species A; microscopic sample image B includes images of cell type B of species B; the cell type A images are labeled to obtain labeled cell type A images of species A; the cell type B images are labeled to obtain labeled cell type B images of species B; AI training is performed using labeled cell images A and B to obtain an AI species cell feature dataset; the AI species cell feature dataset is combined with the corresponding AI algorithm module to have the ability to identify and distinguish cells of species A from cells of species B.
[0013] The aforementioned microscopic sample image A or microscopic sample image B is derived from the detection data, which is a detection image of a suspended microscopic sample.
[0014] Samples containing cells of a selected species are preprocessed to obtain microscopic samples; the microscopic samples are laid flat; the species cells are allowed to suspend and sink to the bottom; the flat microscopic samples are photographed to obtain microscopic sample images; the selected species cells in the microscopic sample images are identified and labeled to obtain labeled images; the labeled images are used for AI training to obtain an AI species cell feature dataset; the AI species cell feature dataset includes the cell features of the selected species samples; the AI species cell feature dataset is combined with the corresponding AI algorithm module to have the ability to identify cells of the selected species.
[0015] The AI training method for identifying the species of cell origin mentioned above includes any one of the following technical features: Feature TC1: Cell type A is a nucleated red blood cell of fish, and cell type B is annulled red blood cell of mammals; Feature TC2: Cell type A is a nucleated red blood cell of birds, and cell type B is annulled red blood cell of mammals; Feature TC3: Cell type A is a nucleated red blood cell of reptiles, and cell type B is annulled red blood cell of mammals; Feature TC4: Cell type A is a nucleated red blood cell of mammals, and cell type B is annulled red blood cell of mammals.
[0016] The AI training method for identifying the species of cell origin mentioned above includes any one of the following technical features: Feature TD1: Cell type A is fish leukocytes and cell type B is mammalian leukocytes; Feature TD2: Cell type A is bird leukocytes and cell type B is mammalian leukocytes; Feature TD3: Cell type A is reptile leukocytes and cell type B is mammalian leukocytes.
[0017] The AI training method for identifying the species of cell origin mentioned above includes any one of the following technical features: Feature TE1: Cell type A is fish coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets; Feature TE2: Cell type A is bird coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets; Feature TE3: Cell type A is reptile coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets.
[0018] The technical solution of this application to solve the above-mentioned technical problem can also be a detection method for identifying the cell-derived species. The blood sample is pre-processed to obtain a microscopic sample; the microscopic sample is laid flat; cells are allowed to suspend and settle at the bottom; the flat microscopic sample is photographed to obtain a microscopic sample image; an AI recognition algorithm is used to identify the microscopic sample image and identify the cell-derived species in the sample within a selected area S1 of the image; the AI recognition algorithm identifies the cell-derived species in the sample based on an AI species cell feature dataset; the AI species cell feature dataset is obtained by training on labeled cell images, and the cell images originate from two or more species.
[0019] The above-mentioned detection method for identifying cell-derived species includes blood-derived species including any two or more of birds, reptiles, fish, and mammals.
[0020] The aforementioned cells include any one of the following: white blood cells, red blood cells, coagulation cells or clusters of coagulation cells, mammalian platelets or agglutinated platelets.
[0021] The above-mentioned detection method for identifying the cell source species includes any one of the following technical features: TF1: Using an AI recognition algorithm to identify the above-mentioned microscopic sample image, identify the total number of cells corresponding to species A (NUMA) and the total number of cells corresponding to species B (NUMB) in the sample within the selected area S1 of the image; the volume of the microscopic sample corresponding to the selected area S1 is V1; the cell unit volume content of species A in the above-mentioned microscopic sample is NUMA / V1; the cell unit volume content of species B in the above-mentioned microscopic sample is NUMB / V1; TF2: Using an AI recognition algorithm to identify the above-mentioned microscopic sample image, identify the total number of cells corresponding to species A (NUMA) and the total number of cells corresponding to species B (NUMB) in the sample within the selected area S1 of the image; the ratio of cells corresponding to species A to cells corresponding to species B = NUMA / NUMB.
[0022] The technical solution of this application to solve the above-mentioned technical problem can also be a computing processing device, including any one of the following technical features: TH1: for running all or part of the above-mentioned method; TH2: the memory of the above-mentioned computing processing device includes all or part of the data of the above-mentioned method.
[0023] The technical solution of this application to solve the above-mentioned technical problems can also be a data storage device, including any one of the following technical features: TG1: storing all or part of the program code for executing the above method; TG2: storing all or part of the data of the above method.
[0024] The technical solution of this application to solve the above-mentioned technical problems can also be a detection device, used to perform part or all of the above-mentioned methods, or to store all or part of the data of the above-mentioned methods.
[0025] The technical effects of the above-mentioned technical solution include: using advanced AI computing power to identify material types based on the cellular characteristics of different species, making it possible to identify species using machines, and making species identification more efficient and accurate.
[0026] The technical effects of the above-mentioned technical solution include: packaging the AI species cell feature dataset, which includes nucleated red blood cell features, obtained through training into independent data that can be sold separately, reducing the engineering difficulty of the entire industry and enabling new technologies to be promoted and implemented as soon as possible.
[0027] The technical effects of the above-mentioned technical solution include: the combination of AI species cell feature dataset and corresponding AI algorithm module has the ability to identify cells of selected species, and can improve the identification efficiency of multiple cell sources, especially when samples from different species are mixed.
[0028] The technical effects of the above-mentioned technical solution include: cell type A includes nucleated red blood cells from fish, birds, reptiles, and mammals, enabling the recognition of nucleated red blood cells from different species.
[0029] The technical effects of the above-mentioned technical solution include: cell type A includes white blood cells from fish, birds, reptiles, and mammals, enabling the recognition of white blood cells from different species.
[0030] The technical effects of the above-mentioned technical solution include: cell type A includes coagulation cells or coagulation cell clusters of fish, birds, reptiles, and mammals, which can realize the ability to identify coagulation cells or coagulation cell clusters of different species.
[0031] The technical effects of the above-mentioned technical solution include: identifying the species of origin of cells in samples through species cell feature datasets, providing efficient support technology for scientific research and reconnaissance, and providing a powerful tool for automatic species identification.
[0032] The technical effects of the above-mentioned technical solution include: it can not only identify species, but also calculate the content and ratio of different cells, providing accurate quantitative reference for analysis. Attached Figure Description
[0033] Figures 1 to 2 These are schematic diagrams 1 and 2 of an AI training method for identifying cell-derived species;
[0034] Figure 3 This diagram illustrates mature red blood cells from different species. Reptiles, birds, fish, and amphibians have nucleated red blood cells, while mammals have anucleated red blood cells.
[0035] Figure 4 It is a schematic diagram of different species and types of red blood cells, including immature nucleated cells, reticulocytes, and shadow cells;
[0036] Figure 5 This is a schematic diagram of coagulation cells and coagulation cell clusters from different species.
[0037] Figure 6 This is a schematic diagram of white blood cells in reptiles, birds, fish, and amphibians;
[0038] Figure 7 This is a schematic diagram of white blood cells in mammals;
[0039] Figure 8 These are schematic diagrams of red blood cells from different species, including schematic diagrams of red blood cells from dogs and cats, including mature anucleate red blood cells and nucleated red blood cells from dogs and cats.
[0040] Figure 9 It shows diagrams of red blood cells and white blood cells in turtles, as well as platelets in turtles and lizards;
[0041] Figure 10 These are schematic diagrams illustrating the identification markers for snake blood cells and turtle blood cells.
[0042] Figures 11 to 13 These are schematic diagrams 1 to 3 illustrating detection methods for identifying cell-derived species. Detailed Implementation
[0043] The contents of this application will be further described in detail below with reference to the accompanying drawings. It should be noted that the following description is of preferred embodiments of the present invention and does not constitute any limitation on the present invention. The description of the preferred embodiments of the present invention is merely an explanation of the general principles of the invention. The designations "first," "second," "A," and "B" used in this invention are for ease of explanation only and do not represent a temporal or spatial order. The combinations of letters and numbers "TA," "TB," and "H" used in this invention are for ease of explanation only, and their specific meanings are determined by the specific terms they represent.
[0044] like Figure 1 In one embodiment of an AI training method for identifying cell origin species, microscopic sample image A includes cell type A images of species A; microscopic sample image B includes cell type B images of species B; cell type A images are labeled to obtain cell type labeled image A of species A; cell type A images are labeled to obtain cell type labeled image B of species B; AI is trained using cell type labeled image A and cell type labeled image B of species B; the AI species cell feature dataset is combined with the corresponding AI algorithm module to have the ability to identify and distinguish cells of species A from cells of species B.
[0045] like Figure 2 The aforementioned microscopic sample image A or microscopic sample image B is derived from the detection data, which is a detection image of a suspended microscopic sample.
[0046] like Figure 2 In one embodiment of an AI training method for identifying the species of cell origin, a sample containing cells of a selected species is preprocessed to obtain a microscopic sample; the microscopic sample is laid flat; the species cells are suspended and allowed to settle at the bottom; the flat microscopic sample is photographed to obtain a microscopic sample image; the selected species cells in the microscopic sample image are identified and labeled to obtain labeled images; the labeled images are used for AI training to obtain an AI species cell feature dataset; the AI species cell feature dataset includes the cell features of the selected species sample; the AI species cell feature dataset is combined with the corresponding AI algorithm module to have the ability to identify cells of the selected species. The selected species cells include any one or more types of red blood cells, white blood cells, and coagulation cells of the corresponding species.
[0047] like Figures 3 to 4 ,as well as Figure 8 and Figure 9In one embodiment of an AI training method for identifying the species of cell origin, cell type A is nucleated red blood cells from fish and cell type B is anucleated red blood cells from mammals; cell type A is nucleated red blood cells from birds and cell type B is anucleated red blood cells from mammals; cell type A is nucleated red blood cells from reptiles and cell type B is anucleated red blood cells from mammals; cell type A is nucleated red blood cells from mammals and cell type B is anucleated red blood cells from mammals. Reptiles include snakes and turtles. Species may also include amphibians. Figure 9 The diagrams of red blood cells and white blood cells from turtles, as well as platelets from turtles and lizards, can be used for AI training that combines them with other cells from other species.
[0048] like Figure 10 In one embodiment of an AI training method for identifying the species of cells, cell type A is a nucleated red blood cell from a snake, and cell type B is a nucleated red blood cell from a turtle. The trained model can simultaneously identify nucleated red blood cells from both snakes and turtles, thus achieving species identification for turtles and snakes.
[0049] like Figures 6 to 7 In one embodiment of an AI training method for identifying the species of cell origin, cell type A is fish leukocytes and cell type B is mammalian leukocytes; cell type A is bird leukocytes and cell type B is mammalian leukocytes; cell type A is reptile leukocytes and cell type B is mammalian leukocytes.
[0050] like Figure 5 In one embodiment of an AI training method for identifying the species of cell origin, cell type A is fish coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets; cell type A is bird coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets; cell type A is reptile coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinating platelets.
[0051] like Figure 11In one embodiment of a detection method for identifying the species of origin of cells, a blood sample is preprocessed to obtain a microscopic sample; the microscopic sample is laid flat; cells are allowed to suspend and settle at the bottom; the flat microscopic sample is photographed to obtain a microscopic sample image; an AI recognition algorithm is used to identify the microscopic sample image and identify the species of origin of cells in the sample within a selected area S1 of the image; the AI recognition algorithm identifies the species of origin of cells in the sample based on an AI species cell feature dataset; the AI species cell feature dataset is obtained by training on labeled cell images, and the cell images originate from two or more species. The blood-originating species include any two or more species from fish, reptiles, amphibians, and mammals. The above classification does not limit specific animal types; animal type classifications can be more numerous and refined. In principle, the classification can be refined to each individual animal. For example, reptiles can include snakes, turtles, crocodiles, etc.
[0052] like Figures 3 to 10 The aforementioned cells include white blood cells, red blood cells, coagulation cells or clusters of coagulation cells, mammalian platelets or agglutinated platelets.
[0053] like Figure 12 In one embodiment of a detection method for identifying the species of cell origin, an AI recognition algorithm is used to identify the microscopic sample image and determine the total number of cells corresponding to species A (NUMA) and species B (NUMB) in the sample within a selected area S1 of the image. The volume of the microscopic sample corresponding to the selected area S1 is V1; the cell volume content per unit volume of species A in the microscopic sample is NUMA / V1; and the cell volume content per unit volume of species B in the microscopic sample is NUMB / V1. The cell volume content per unit volume of species A is equivalent to the cell volume content of the species.
[0054] like Figure 13 In one embodiment of a detection method for identifying the species of cell origin, an AI recognition algorithm is used to identify the microscopic sample image and to identify the total number of cells corresponding to species A (NUMA), the total number of cells corresponding to species B (NUMB), and the ratio of cells corresponding to species A to cells corresponding to species B in the sample within a selected area S1 of the image.
[0055] =NUMA / NUMB.
[0056] Figure 10 The left side shows a diagram of the identification markers for snake blood cells, and the right side shows a diagram of the identification markers for turtle blood cells. Figure 10 It includes nucleated red blood cells and white blood cells.
[0057] A computing processing apparatus is used to run all or part of the above-described methods; the memory of the computing processing apparatus includes all or part of the data of the above-described methods.
[0058] A data storage device stores all or part of the program code for performing the above methods; and stores all or part of the data for performing the above methods.
[0059] A detection device for performing part or all of the above methods, or storing all or part of the data of the above methods.
[0060] While the invention has been described and illustrated with reference to preferred embodiments and several alternatives, it is not intended to be limited to the specific descriptions herein. Other alternatives or equivalent components may also be used to practice the invention.
Claims
1. An AI training method for identifying cell-derived species, characterized in that, The microscopic sample image A includes images of cell type A of species A; The microscopic sample image B includes images of cell types B of species B; Label the image of cell type A to obtain cell type labeled image A of species A; Image B is labeled with cell type B to obtain cell type labeled image B of species B; AI was trained using cell-labeled image A and cell-labeled image B to obtain an AI species cell feature dataset. The AI species cell feature dataset, combined with the corresponding AI algorithm module, has the ability to identify and distinguish cells of species A from cells of species B.
2. The AI training method for identifying cell-derived species according to claim 1, characterized in that, The microscopic sample image A or microscopic sample image B is derived from the detection data. The detection data are detection images obtained from microscopic imaging of suspended microscopic samples.
3. The AI training method for identifying cell-derived species according to claim 1, characterized in that, Samples containing cells of selected species are pretreated to obtain microscopic samples; The microscopic sample is laid flat; the cells of the species are allowed to suspend and settle at the bottom. Take pictures of the laid-out microscopic samples to obtain microscopic images of the samples; Selected species cells in microscopic sample images are identified and labeled to obtain labeled images. AI is then trained using these labeled images to obtain an AI species cell feature dataset. The AI species cell feature dataset includes cell features of selected species samples; The AI species cell feature dataset, combined with the corresponding AI algorithm module, has the ability to identify cells of selected species.
4. The AI training method for identifying cell-derived species according to claim 1, characterized in that, Includes any one of the following technical features: Characteristic TC1: Cell type A is a nucleated red blood cell of fish, and cell type B is an anucleate red blood cell of mammals; Characteristic TC2: Cell type A is nucleated erythrocytes of birds, and cell type B is anucleated erythrocytes of mammals; Characteristic TC3: Cell type A is a nucleated erythrocyte of reptiles, and cell type B is annulled erythrocyte of mammals; Characteristic TC4: Cell type A is a nucleated erythrocyte of mammals, and cell type B is annulled erythrocyte of mammals.
5. The AI training method for identifying cell-derived species according to claim 1, characterized in that, Includes any one of the following technical features: Feature TD1: Cell type A is fish leukocytes, cell type B is mammalian leukocytes; Feature TD2: Cell type A is avian leukocytes, cell type B is mammalian leukocytes; Feature TD3: Cell type A is a reptilian leukocyte, and cell type B is a mammalian leukocyte.
6. The AI training method for identifying cell-derived species according to claim 5, characterized in that, Includes any one of the following technical features: Feature TE1: Cell type A is fish coagulation cells or coagulation cell clusters; cell type B is mammalian platelets or agglutinated platelets. Characteristic TE2: Cell type A is avian coagulation cells or coagulation cell clusters; cell type B is mammalian platelets or agglutinated platelets. Characteristic TE3: Cell type A is reptilian coagulation cells or coagulation cell clusters, and cell type B is mammalian platelets or agglutinated platelets.
7. A detection method for identifying cell-derived species, characterized in that, Blood samples are pretreated to obtain microscopic samples; Spread the microscopic sample flat; allow the cells to suspend and settle at the bottom. Take pictures of the laid-out microscopic samples to obtain microscopic images of the samples; The AI recognition algorithm is used to identify the microscopic sample image and to identify the cell-originating species in the sample within a selected area S1 of the image; the AI recognition algorithm identifies the cell-originating species in the sample based on an AI species cell feature dataset. The AI species cell feature dataset is obtained by training with labeled cell images, which are from two or more species.
8. The detection method for identifying cell-derived species according to claim 7, characterized in that, The blood source species include any two or more of the following: birds, reptiles, fish, and mammals.
9. The detection method for identifying cell-derived species according to claim 8, characterized in that, The cells include any one of white blood cells, red blood cells, coagulation cells or clusters of coagulation cells, mammalian platelets or agglutinated platelets.
10. The detection method for identifying cell-derived species according to claim 8, characterized in that, Includes any one of the following technical features: TF1: Use an AI recognition algorithm to identify the microscopic sample image and identify the total number of cells corresponding to species A (NUMA) and species B (NUMB) in the sample within the selected area S1 of the image; the volume of the microscopic sample corresponding to the selected area S1 is V1; the cell unit volume content of species A in the microscopic sample is NUMA / V1; the cell unit volume content of species B in the microscopic sample is NUMB / V1. TF2: Use AI recognition algorithm to identify the microscopic sample image and identify the total number of cells corresponding to species A (NUMA) and the total number of cells corresponding to species B (NUMB) in the sample within the selected area S1 of the image; the ratio of cells corresponding to species A to cells corresponding to species B = NUMA / NUMB.
11. A computing processing device, characterized in that, Includes any one of the following technical features: TH1: Used to perform all or part of the method according to any one of claims 1 to 10; TH2: The memory of the computing processing device includes all or part of the data of the method according to any one of claims 1 to 10.
12. A data storage device, characterized in that, Includes any one of the following technical features: TG1: Stores all or part of the program code for performing the method according to any one of claims 1 to 10; TG2: Data containing all or part of the method according to any one of claims 1 to 10.
13. A detection device, characterized in that, Used to perform part or all of the method described in any one of claims 1 to 10 Or store all or part of the data of the method according to any one of claims 1 to 10.