Blood parasite detection and ai training method and detection sample containment device

Abstract

The blood parasite detection and AI training method and the detection sample containing device, collect blood of parasite infected animals; the blood is pretreated to obtain a detection sample; the blood pretreatment includes dilution treatment of the blood with a diluent; the blood parasites are allowed to assume a natural state in the liquid; the detection sample is photographed to obtain a detection sample image; the blood parasites in the detection sample image are identified and labeled to obtain a labeled picture, which is used for AI training to obtain a blood parasite AI feature data set; the blood parasite AI feature data set is combined with a corresponding AI algorithm to have the ability to identify blood parasites. The blood parasite detection and the detection sample containing device are used for detecting blood parasites.

Description

Technical Field

[0001] This application belongs to the field of formed element analysis technology based on magnified microscopic images, and specifically relates to a blood parasite infection assessment parameter, a blood parasite detection method, and its AI training method, computational processing, and storage device. Background Technology

[0002] Bloodworms. According to Chai Jianyuan et al. (1990), six species of bloodworms have been found in turtles and tortoises in my country. Taking *Hydroceta sinensis* as an example, it exists in the Chinese soft-shelled turtle in three stages:

[0003] ① Mature schizonts within Kupffer cells of the liver are oval or spherical, measuring 17.92-18.24 μm × 14.84-15.24 μm, and produce 15-20 merozoites, which are short rod-shaped.

[0004] ② In deep blood, schizont proliferation occurs within the erythrocytes of the hepatic sinusoids. There are X-type and Y-type schizonts, both similar in size and shape, and both columnar. X-type schizonts produce 14-18 small, wedge-shaped merozoites; Y-type schizonts produce 4-8 large, rod-shaped merozoites. Female gametocytes are kidney-shaped, while male gametocytes are comma-shaped with a small tail at the posterior end.

[0005] ③ In peripheral blood cells, the early-stage schizont is broadly kidney-shaped and produces 6-12 merozoites, which are banana-shaped; the intermediate-stage schizont produces 6-10 merozoites; and the late-stage schizont produces 2 merozoites. Gametocytes originate from Y-type merozoites, and the vegetative body is the intermediate stage in the development from the previous generation of merozoites to the next generation. The Chinese soft-shelled turtle is the intermediate host of *Hydrocotyle sinensis*.

[0006] The leech *Hirudo medicinalis* is the definitive host of *Hydrocotyle sinensis*. After the leech feeds on the blood of a diseased turtle, *Hydrocotyle sinensis* develops within the leech's digestive tract, involving two stages: gametogony and spore reproduction. Gametogony is characterized by the fusion of male and female gametoblast mother cells before gamete differentiation, producing four male gamete nuclei. One of these nuclei fertilizes the female gamete nucleus, forming a zygote. Spore reproduction begins with nuclear division of a mononuclear oocyst, eventually forming a mature oocyst containing eight gymnosporidis, which then disintegrates to release sporospores.

[0007] Canine and Feline Hematology: Hepatic tufts. Hepatic tufts are a type of specific intracellular parasitic protozoan belonging to the family Hepatidae. More than 340 species of hepatic tufts have been described, found not only in mammals but also in invertebrates, reptiles, birds, and vertebrate marsupials. The first report of a domestic cat dates back to 1908, when it was called *Leucocytozoonfelis domestici*, and was later reclassified as a hepatic tuft. Similar to the species that infect dogs, feline hepatic tufts are considered a dominant species in cat infections; however, there is also evidence of canine hepatic tufts infecting cats.

[0008] Canine liver tuft infection is widespread in Southern Europe, Africa, Asia, and South America. Dogs with low-level infection may not show clinical symptoms, while dogs with high-level infection may be fatal when accompanied by symptoms such as fever, lethargy, and anemia.

[0009] A long, thick, rod-shaped pathogen can be seen inside a neutrophil in a cat blood smear; this is a gametoblast of *Hepatocytozoa felis*.

[0010] In a canine blood smear, a long, thick, rod-shaped pathogen is visible in the center of the field of view within a neutrophil; this is a gametoblast cell of *Hepatocytozoa canis*.

[0011] The applicant has filed a series of Chinese patents, such as

[0012] 1. CN2020112669290, "Cellular analysis methods and systems and quantitative methods and systems";

[0013] 2. CN2020112669182, "Imaging methods, systems and kits for cell suspension samples";

[0014] 3. CN2022104799126, "A method for rapid focusing of a microscopic image acquisition device and a method for acquiring microscopic images";

[0015] 4. CN2023110496905, "A device, chip and method for detecting formed elements in blood, urine and feces (multi-parameter)".

[0016] A brand-new technical solution is used to measure the content of target substances in blood.

[0017] The applicant submitted Chinese patent "2024104905675", entitled "Method for Detection of Fecal Parasites and AI Training and Infection Assessment Parameters", which proposed a method for detecting fecal parasites based on AI image recognition. However, blood parasites, most of which live in blood cells and are mixed with substances in the cell nucleus, have small images and indistinct features, making them difficult for AI to distinguish. How to detect parasites in blood is a completely new technical problem. Summary of the Invention

[0018] Through extensive experimentation and AI training, the applicant discovered that AI can automatically extract images of blood parasites, particularly under high-intensity white light or ultraviolet light, clearly highlighting the features of parasites within cells. The AI's recognition accuracy reaches the level required for medical testing. By learning from AI images of parasites at different stages, it can, to a certain extent, distinguish different stages of infection. This represents a pioneering new detection technology in the field of blood parasite detection.

[0019] The technical solution proposed in this application is a method for training AI to detect blood parasites, comprising: collecting blood from animals infected with parasites; pre-processing the blood to obtain test samples; the blood pre-processing includes diluting the blood with a diluent; allowing the blood parasites to exist in their natural state in the liquid; photographing the test samples to obtain test sample images; identifying and labeling the blood parasites in the test sample images to obtain labeled images; using the labeled images for AI training to obtain a blood parasite AI feature dataset; and combining the blood parasite AI feature dataset with the corresponding AI algorithm to achieve the ability to identify blood parasites.

[0020] The test sample is illuminated with white light, and the intensity of the white light is required to clearly show the image of the material inside the cell nucleus. The test sample is then photographed to obtain an image of the test sample. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to have the ability to identify blood parasites in the blood parasite images collected from the test sample illuminated with white light.

[0021] The test sample is illuminated with ultraviolet light, and an image of the test sample is obtained. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to identify blood parasites in the blood parasite images collected from the test sample illuminated with ultraviolet light.

[0022] The AI ​​training method for blood parasite detection includes any one of the following technical features: TF10: the parasite is *Hematoxylin and cystis*; TF20: the infected animal is a reptile, the parasite is *Hematoxylin and cystis*, and it resides in red blood cells; TF30: the infected animal is an avian, the parasite is *Hematoxylin and cystis*, and it resides in red blood cells; TF40: the infected animal is an amphibian, the parasite is *Hematoxylin and cystis*, and it resides in red blood cells; TF50: the infected animal is a mammal, the parasite is *Hematoxylin and cystis*, and it resides in red blood cells; TF60: the infected animal is a mammal, the parasite is *Hematoxylin and cystis*, and it resides in white blood cells.

[0023] The AI ​​training method for detecting blood parasites includes any one of the following technical features: TK10: The infected animal is a reptile, which is infected by injecting the blood of an animal infected with blood clostridial parasites; TK20: The infected animal is a bird, which is infected by injecting the blood of an animal infected with blood clostridial parasites; TK30: The infected animal is an amphibian, which is infected by injecting the blood of an animal infected with blood clostridial parasites.

[0024] The AI ​​training method for blood parasite detection includes: collecting blood samples from different infection stages after infection of experimental animals; the blood parasite feature dataset includes a parasite infection stage feature dataset; the infection stage is expressed in days or hours.

[0025] The AI ​​training method for blood parasite detection includes any one of the following technical features: TN10: Blood pretreatment includes staining of the test sample, in which a staining agent is added to the blood; TN20: Blood pretreatment includes staining of the test sample, in which the staining agent is dissolved in a diluent beforehand.

[0026] Normal red blood cells in the test sample images are identified and labeled to obtain labeled images. AI is then trained using these labeled images to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to enable the identification of normal red blood cells.

[0027] White blood cells in the sample images are identified and labeled to obtain labeled images. AI is then trained using these labeled images to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to enable the identification of white blood cells.

[0028] The technical solution to the technical problem addressed in this application can also be a blood parasite detection method, which involves pre-processing blood to obtain a test sample; the blood pre-processing includes diluting the blood with a diluent; allowing the blood parasites to remain in the liquid in their natural state; photographing the test sample to obtain an image of the test sample; combining a blood parasite AI feature dataset with a corresponding AI algorithm to identify the test sample image and identify the parasites in the image; the blood parasite AI feature dataset is obtained by AI training on labeled blood parasite images.

[0029] The test sample is illuminated with white light, and the sample is photographed. The intensity of the white light needs to be sufficient to clearly show the image of the material inside the cell nucleus. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to identify blood parasites in the blood parasite images collected from the test sample illuminated with white light.

[0030] The test sample is illuminated with ultraviolet light, and an image of the test sample is obtained. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to identify blood parasites in the blood parasite images collected from the test sample illuminated with ultraviolet light.

[0031] The AI ​​recognition algorithm is used to identify the sample image and identify the blood parasites in the sample within the selected area S1 of the image, thus obtaining the total number of blood parasites NUM1 in the sample of the selected image.

[0032] The volume of the sample corresponding to the selected area S1 is V1, and the volume content of blood parasites in the sample is NUM1 / V1.

[0033] The AI ​​feature dataset for blood parasites was obtained by training images labeled with manual annotations of the parasite infection stages, or by training blood images obtained from experimental animals after N days of infection.

[0034] The blood parasite detection method includes any one of the following technical features: TH10: output the number of parasites at different infection stages; TH20: output the total number of selected parasites; TH30: output the percentage of parasites in different infection periods; the blood parasite feature dataset is obtained by training labeled blood parasite images.

[0035] The method for detecting blood parasites includes any one of the following technical characteristics: TG10: the parasite is *Hematoxylin and cystis*; TG20: the infected animal is a reptile, the parasite is *Hematoxylin and cystis*, and it resides within red blood cells; TG30: the infected animal is an avian, the parasite is *Hematoxylin and cystis*, and it resides within red blood cells; TG40: the infected animal is an amphibian, the parasite is *Hematoxylin and cystis*, and it resides within red blood cells; TG50: the infected animal is a mammal, the parasite is *Hematoxylin and cystis*, and it resides within red blood cells; TG60: the infected animal is a mammal, the parasite is *Hematoxylin and cystis*, and it resides within white blood cells.

[0036] The AI ​​recognition algorithm is used to identify the sample image and identify the cells of the selected category in the sample within the selected area S1 of the image, and obtain the total number of cells of the selected category NUM3 in the sample of the selected image; the AI ​​recognition algorithm identifies the cells of the selected category in the sample based on the cell feature dataset of the selected category; the cell feature dataset of the selected category is obtained by training on the labeled cells of the selected category.

[0037] The method for detecting blood parasites includes any one of the following technical features: TQ10: The selected cell types include any one or more of coagulation cells, coagulation cell clusters, leukocytes, immature nucleated cells, reticulocytes, and shadow cells; TQ20: The content of the selected cell type per unit volume of the sample is NUM3 / V1; TQ30: The ratio of blood parasites to the number of selected cell types is calculated as NUM1 / NUM3.

[0038] A method for detecting blood parasites involves using an AI recognition algorithm to identify the sample image, pinpointing normal red blood cells within a selected area S1 of the image, and obtaining the total number of normal red blood cells NUM2 in the selected image. The AI ​​recognition algorithm identifies normal red blood cells in the sample based on a dataset of normal red blood cell characteristics. The ratio of blood parasites to normal red blood cells is calculated as NUM1 / NUM2. The samples can be either mammalian or non-mammal.

[0039] In this application, the inventors propose a device and a sample container specifically for detecting blood parasites; the detection of blood parasites is performed using an image AI computing module.

[0040] The technical solution to the technical problem addressed in this application can also be a blood parasite detection device, used to detect blood parasites. The blood undergoes pretreatment to obtain a test sample, which is then added to the receiving cavity of a test sample receiving device. The device includes a control module, a camera assembly, a support assembly, and an image AI computing module. The control module is electrically connected to the camera assembly; the control module is also electrically connected to the image AI computing module; the support assembly is used to place the test sample receiving device; the image AI computing module includes a storage unit for storing blood parasite AI feature data; the camera assembly is used to capture images of the test sample to obtain test images; and the image AI computing module is used to analyze the test images.

[0041] A blood parasite detection device includes any one or more of the following technical features: TA10: an image AI computing module is located within a control module; TA20: the control module includes a network component; the image AI computing module is located in a server, and the network component is electrically connected to the image AI computing module via a network.

[0042] The camera assembly includes a Z-axis slide assembly and a microscope camera assembly. The microscope camera assembly is mechanically connected to the slide in the Z-axis slide assembly, and the microscope camera assembly can move in the Z-axis direction.

[0043] The blood parasite detection device also includes a drive module, a control module electrically connected to the drive module, and a drive module electrically connected to the support component. The support component includes an X-slide assembly and a sample receiving device placed on the slide of the X-slide assembly. The drive module is used to drive the X-slide assembly to move in the X-axis direction, thereby moving the sample receiving device in the X-axis direction.

[0044] The blood parasite detection device includes a support component consisting of a Y-stage assembly and an X-stage assembly placed on the slide of the Y-stage assembly. A drive module is used to drive the Y-stage assembly to move in the Y-axis direction, thereby moving the sample container in the X-axis or Y-axis direction.

[0045] A blood parasite detection device includes any one of the following technical features: TB10: further includes circuit board A and circuit board B, with the control module disposed on circuit board A and the drive module disposed on circuit board B; TB20: further includes circuit board C, with the drive module and the control module disposed on circuit board C.

[0046] A blood parasite detection device, comprising any one or more of the following technical features: TC10: The support assembly includes a light source assembly, the light source assembly including a white light source, the white light emitted by the white light source being used to illuminate the receiving cavity; TC20: The support assembly includes a light source assembly, the light source assembly including a purple light source, the purple light emitted by the purple light source being used to illuminate the receiving cavity; TC30:

[0047] The camera component is located above or below the support component.

[0048] The technical solution to the technical problem solved by this application can also be a sample containing device for blood parasite detection. The blood is pretreated to obtain a test sample, which is then added to the containing cavity of the sample containing device, including containing cavity A, containing cavity B, and a sample inlet. The sample inlet is used to add the test sample to be tested. The sample inlet is connected to containing cavity A and containing cavity B. The liquid containing height of containing cavity A is higher than that of containing cavity B. Containing cavity A is used to contain the test sample of species A, and containing cavity B is used to contain the test sample of species B.

[0049] A sample containment device includes any one or more of the following technical features: D10: Containment chambers A and B are connected in series for sequentially filling the liquid to be tested into containment chambers A and B; TD20: Containment chambers A and B are connected in parallel for simultaneously filling the liquid to be tested into containment chambers A and B; TD30: Containment chambers A and B are connected in parallel but not connected for separately filling the liquid to be tested into containment chambers A and B; TD40: It also includes an exhaust port; the exhaust port is connected to containment chamber A; the exhaust port is connected to containment chamber B; the exhaust port is connected to the external atmosphere.

[0050] The sample containment device includes any one of the following technical features: Feature TE10: the upper part of containment cavity A and the upper part of containment cavity B are made of transparent material; Feature TE20: the lower part of containment cavity A and the lower part of containment cavity B are made of transparent material; Feature TE30: containment cavity A and containment cavity B have the same bottom height; Feature TE40: containment cavity A and containment cavity B have the same top height.

[0051] The technical effects of the solution include: by collecting blood from parasite-infected animals and artificially enriching it for training, the concentration of the target substance in the sample can be significantly increased, thereby improving the efficiency of AI training.

[0052] The technical benefits of this solution include: Evolution based on a blood parasite feature dataset allows for iterative improvement of AI recognition capabilities. Continuous training of the AI ​​recognition model enriches the dataset and increases recognition accuracy. Iterative model updates create a closed-loop feedback loop, enabling the instrument to continuously improve its recognition capabilities during daily use.

[0053] The technical benefits of this solution include: the white light intensity is sufficient to clearly image the substances within the cell nucleus, enabling clear identification of parasites within cells based on strong white light. Training with images acquired under white light results in higher recognition accuracy for images acquired under the same conditions.

[0054] The technical benefits of this solution include: Illuminating the sample with ultraviolet light makes it easier to identify and distinguish intracellular parasites and other substances within the cell nucleus; in particular, the absorption characteristics of ultraviolet light by hemoglobin in red blood cells differ from those of parasites, thus ultraviolet light irradiation naturally increases the image's distinguishability and improves recognition accuracy. Training with images acquired under ultraviolet light results in higher recognition accuracy for images acquired under the same conditions.

[0055] The technical advantages of the solution include: it is applicable to the detection of blood clotworms in a variety of different types of animals.

[0056] The technical benefits of the solution include: the amount of infection in laboratory animals can be precisely controlled, further facilitating the acquisition of target parasite species and improving training efficiency.

[0057] The technical effects of the solution include: collecting blood samples from different stages of infection to obtain a dataset of parasite infection stage features; and using AI recognition algorithms to identify parasites in images that have infection time dimension features based on the parasite infection stage feature dataset.

[0058] The technical benefits of the solution include: the blood parasite feature dataset includes parasite characteristic data at different stages of infection; the distinction between infection stages allows for deeper parasite analysis, providing clinicians with information on the time dimension of infection. The integration of multi-dimensional information such as infection time, type, and quantity provides clinicians with more accurate reference information.

[0059] The technical advantages of this solution include: simultaneous dilution and staining effectively controls the concentration and uniformity of the staining agent; stained parasites are easier to identify; and the pre-dissolved staining agent in the diluent increases sample preparation efficiency.

[0060] The technical effects of the solution include: combining the blood parasite AI feature dataset with the corresponding AI algorithm, it has the ability to identify normal red blood cells and can simultaneously identify normal red blood cells and cells infected with parasites.

[0061] The technical effects of the solution include: combining the blood parasite AI feature dataset with the corresponding AI algorithm, it has the ability to identify white blood cells, and can simultaneously identify normal white blood cells and cells infected with parasites.

[0062] The technical benefits of the solution include: obtaining the blood parasite content per unit volume as NUM1 / V1, and providing accurate quantitative analysis results.

[0063] The technical benefits of this solution include: controlled culture and phased identification of parasites at various growth stages, yielding deeper information about the parasites. It outputs the number of parasites at different infection stages; the total number of selected parasites; and the percentage of parasites present at different infection time periods, providing multi-dimensional information on parasite infection. The percentage of parasites present at each infection time period = number of parasites present at that time period / total number of parasites, providing another dimension of parasite infection information that can be used to assess the severity of blood parasite infection.

[0064] The technical benefits of the solution include: statistical analysis of parasite infection information based on blood parasite infection assessment parameters, which can output more multi-dimensional analytical data for clinical reference.

[0065] The technical benefits of this solution include: identifying parasites as either bloodworms or liverworms, and recognizing intracellular parasites, thus deepening the understanding of parasite identification. Previously, parasite identification was typically limited to extracellular parasites; identifying intracellular parasites required unconventional microscopic devices, such as electron microscopes and other high-resolution imaging systems, and even then, accurate quantification was difficult.

[0066] The technical benefits of this solution include: the ability to identify selected cell categories.

[0067] The technical benefits of this solution include: allowing blood parasites to exist in a natural state in the liquid; enabling their morphological features to fully unfold, thus forming effective samples for training and achieving sufficient recognition rates. Without naturally unfolded samples, even manual annotation cannot be standardized, hindering AI training. Furthermore, by photographing the test samples and obtaining images corresponding to the same detection application scenario, the AI-trained samples exhibit high recognition rates.

[0068] The technical benefits of this solution include: the ability to collect iterative data on blood parasite characteristics; and the significant improvement in parasite identification efficiency through AI training and application. By applying AI to the analysis of formed elements in blood, AI can replace manual interpretation, greatly increasing efficiency.

[0069] The technical benefits of the solution include: Improved accuracy and consistency through AI recognition: Well-trained AI models can achieve high accuracy and stability, reducing the risk of misdiagnosis and missed diagnosis. AI will not miss tiny eggs, oocysts, cysts, or insect bodies, or misidentify other substances as eggs, oocysts, cysts, or insect bodies due to human visual errors, experience differences, or distraction. Increased efficiency: AI algorithms can process large numbers of images in a short time, unaffected by fatigue or emotional factors, improving detection speed and efficiency, and effectively solving the detection bottleneck caused by time-consuming manual interpretation. Enhanced precision and sensitivity: AI algorithms can identify minute details and subtle color differences, and may be more sensitive to certain types of eggs, oocysts, cysts, or insect bodies with large morphological variations that are difficult to detect.

[0070] The technical benefits of the solution include: the setup of the control module, camera component, support component, and image AI computing module; a dedicated detection device for occasional blood parasites; and the ability to identify blood parasite-infected cells from a massive number of cells using AI, greatly improving identification efficiency.

[0071] The technical benefits of the solution include: the image AI computing module is integrated into the control module, making the device more efficient and compact.

[0072] The technical benefits of the solution include: the control module includes a network component; the image AI computing module is deployed in the server, and the network component is connected to the image AI computing module via electrical signals through the network, which facilitates real-time updates of the image AI computing module and promotes its evolution.

[0073] The technical benefits of the solution include: the camera assembly comprises a Z-axis slide assembly and a microscope camera assembly; the microscope camera assembly is mechanically connected to the slide in the Z-axis slide assembly, allowing the microscope camera assembly to move in the Z-axis direction. This facilitates image acquisition, provides a device for precisely adjusting the imaging-related distance, and makes it easier to obtain clear images.

[0074] The technical benefits of this solution include: the drive module drives the X-slide assembly to move along the X-axis, thereby moving the sample receiving device along the X-axis. The drive module also drives the Y-slide assembly to move along the Y-axis, thereby moving the sample receiving device along either the X-axis or Y-axis. This movement of the sample receiving device along either the X-axis or Y-axis facilitates the acquisition of images from different positions, improving image acquisition and analysis efficiency.

[0075] The technical benefits of the solution include: the control module is located on circuit board A, and the drive module is located on circuit board B. This separate setup helps reduce maintenance costs.

[0076] The technical benefits of the solution include: the drive module and control module are located on circuit board C, making the device more compact and reducing overall cost.

[0077] The technical effects of the solution include: the light source component includes a white light source, the white light emitted by the white light source is used to illuminate the containment cavity, and the strong white light can provide multiple spectra, which is beneficial for the identification of bloodworms.

[0078] The technical effects of the solution include: the purple light emitted by the purple light source is used to illuminate the cavity; the purple light source is particularly beneficial for the identification of bloodworms in cells, especially for some animals whose red blood cells are nucleated red blood cells, the purple light source is very helpful in distinguishing the nucleus of the red blood cell from the bloodworm.

[0079] The technical benefits of the solution include: the camera component can be positioned above or below the support component, allowing for both vertical and horizontal imaging, thus providing greater flexibility in image acquisition.

[0080] The technical benefits of the solution include: the sample container has two chambers, A and B, which can further improve the detection efficiency of blood parasites.

[0081] The technical benefits of the solution include: different connection and communication methods for cavity A and cavity B, adapting to various application scenarios and improving efficiency.

[0082] The technical advantages of the solution include: the upper part of cavity A and the upper part of cavity B are made of transparent material, and the lower part of cavity A and the lower part of cavity B are made of transparent material, which facilitates the application of vertical imaging methods.

[0083] The technical effects of the solution include: the accommodating cavity A and the accommodating cavity B have the same bottom or top height, so that even if the cavity height of accommodating cavity A is different from that of accommodating cavity B, imaging can be easily performed with the bottom or top as a reference, and they have a common imaging reference plane. Attached Figure Description

[0084] Figure 1This is a diagram illustrating the AI ​​training method for blood parasite detection. Figure 1 ;

[0085] Figure 2 This is an illustration of collecting blood from animals infected with parasites. Figure 1 ;

[0086] Figure 3 Image 1 shows the detection of blood clots.

[0087] Figure 4 yes Figure 3 A magnified view of a portion of the image;

[0088] Figure 5 Image 1 shows the detection of blood clots.

[0089] Figure 6 yes Figure 5 A magnified view of a portion of the image;

[0090] Figures 7 to 10 This is a diagram illustrating the methods for treating blood parasites. Figures 1 to 4 ;

[0091] Figures 11 to 16 This is a schematic diagram of a blood parasite detection device. Figures 1 to 6 ;

[0092] Figures 17 to 20 This is a schematic diagram of a sample container. Figures 1 to 4 ; Detailed Implementation

[0093] 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.

[0094] like Figures 1 to 6 A method for training an AI to detect blood parasites includes: collecting blood from animals infected with parasites; pre-processing the blood to obtain a test sample; the blood pre-processing includes diluting the blood with a diluent; allowing the blood parasites to exist in their natural state in the liquid; photographing the test sample to obtain an image of the test sample; identifying and labeling the blood parasites in the test sample image to obtain labeled images; using the labeled images for AI training to obtain a blood parasite AI feature dataset; and combining the blood parasite AI feature dataset with the corresponding AI algorithm to achieve the ability to identify blood parasites.

[0095] In some embodiments, the test sample is illuminated with white light, the intensity of which needs to be sufficient to clearly visualize the material within the cell nucleus. An image of the test sample is then captured. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to enable the identification of blood parasites in images of blood parasites collected under white light illumination. For example... Figure 3 These are images of blood-borne parasites obtained by a blood parasite detection device under white light illumination. Figure 4 yes Figure 3 A magnified diagram of a portion of the image. The requirement for sufficient white light intensity to clearly visualize the material within the cell nucleus means being able to distinguish parasites.

[0096] In some embodiments, the test sample is irradiated with ultraviolet light and photographed to obtain an image of the test sample; the blood parasite AI feature dataset is combined with the corresponding AI algorithm to have the ability to identify blood parasites in the blood parasite images collected from the test sample irradiated with ultraviolet light. Figure 5 This is a detection image of blood-borne parasites obtained by a blood parasite detection device; Figure 6 yes Figure 5 A magnified view of a portion of the image.

[0097] In some embodiments, the parasite is a blood fluke or a liver fluke. In some embodiments, the animal is a reptile, and the parasite is a blood fluke that lives inside red blood cells. Figures 3 to 6 The text describes a turtle blood parasite. In some embodiments, the animal is a bird, and the parasite is a blood worm that lives within red blood cells. In some embodiments, the animal is an amphibian, and the parasite is a blood worm that lives within red blood cells. In some embodiments, the animal is a mammal, and the parasite is a blood worm that lives within red blood cells. In some embodiments, the animal is a mammal, and the parasite is a liver worm that lives within white blood cells.

[0098] like Figure 2 The AI ​​training method for detecting blood parasites involves infecting reptiles by injecting their blood into animals infected with blood clostridial parasites.

[0099] In some embodiments, the infected animal is a bird, which is infected by injecting the blood of an animal infected with the blood clostridial parasite. In some embodiments, the infected animal is an amphibian, which is infected by injecting the blood of an animal infected with the blood clostridial parasite.

[0100] like Figure 2 The AI ​​training method for blood parasite detection involves selecting experimental animals and collecting blood samples after the animals are infected. Blood samples from different infection stages can be collected. The blood parasite feature dataset includes a parasite infection stage feature dataset, with infection stages measured in days or hours.

[0101] In some embodiments, blood pretreatment includes staining of the test sample, which involves adding a staining agent to the blood. In other embodiments, blood pretreatment includes staining of the test sample, where the staining agent is pre-dissolved in a diluent.

[0102] like Figure 3 and Figure 5 The system identifies and labels normal red blood cells in the sample images to obtain labeled images. These labeled images are then used for AI training to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset, combined with the corresponding AI algorithm, has the ability to identify normal red blood cells.

[0103] In some embodiments, white blood cells in the sample images are identified and labeled to obtain labeled images. These labeled images are then used for AI training to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to enable the identification of white blood cells.

[0104] like Figure 7 A method for detecting blood parasites involves pre-processing blood to obtain a test sample; the blood pre-processing includes diluting the blood with a diluent; allowing the blood parasites to remain in a natural state in the liquid; photographing the test sample to obtain an image of the test sample; and combining a blood parasite AI feature dataset with a corresponding AI algorithm to identify the parasites in the test sample image; the blood parasite AI feature dataset is obtained by AI training on labeled blood parasite images.

[0105] In some embodiments, the test sample is illuminated with white light and photographed. The intensity of the white light needs to be sufficient to clearly show the image of the material inside the cell nucleus, thus obtaining an image of the test sample. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to have the ability to identify blood parasites in the blood parasite images collected by illuminating the test sample with white light.

[0106] In some embodiments, the test sample is illuminated with ultraviolet light and photographed to obtain an image of the test sample; the blood parasite AI feature dataset is combined with the corresponding AI algorithm to have the ability to identify blood parasites in the blood parasite images collected from the test sample illuminated with ultraviolet light.

[0107] like Figures 7 to 8 A method for detecting blood parasites uses an AI recognition algorithm to identify blood parasites in a selected area S1 of the image, obtaining the total number of blood parasites NUM1 in the selected image. The volume of the sample corresponding to the selected area S1 is V1, and the concentration of blood parasites per unit volume of the sample is NUM1 / V1. Figure 8If a blood sample is diluted by N times, the volume can be calculated using V1*N, and then NUM1 / (V1*N) can be used to convert it into the content of blood parasites per unit volume in the original blood sample.

[0108] The AI ​​feature dataset for blood parasites was obtained by training images labeled with manual annotations of the parasite infection stages, or by training blood images obtained from experimental animals after N days of infection.

[0109] The blood parasite detection method can output the number of parasites at different stages of infection; TH20: outputs the total number of selected parasites. The blood parasite detection method can also output the percentage of parasites at different infection time periods; the blood parasite feature dataset is obtained by training with labeled blood parasite images.

[0110] Methods for detecting blood parasites: The parasite is either a blood clotridae or a liver clotridae; the animal is a reptile, and the parasite is a blood clotridae that lives inside red blood cells; the animal is a bird, and the parasite is a blood clotridae that lives inside red blood cells; the animal is an amphibian, and the parasite is a blood clotridae that lives inside red blood cells; the animal is a mammal, and the parasite is a blood clotridae that lives inside red blood cells.

[0111] like Figure 10 The AI ​​recognition algorithm identifies the selected cell types in the sample image within a selected area S1 of the image, obtaining the total number of cells of the selected cell type NUM3 in the sample of the selected image. The AI ​​recognition algorithm identifies cells of the selected cell type in the sample based on the cell feature dataset of the selected cell type. The cell feature dataset of the selected cell type is obtained by training on labeled cells of the selected cell type. The selected cell type includes any one or more of the following: coagulation cells, coagulation cell clusters, leukocytes, immature nucleated cells, reticulocytes, and shadow red blood cells; TQ20: the content of the selected cell type per unit volume of the sample is NUM3 / V1; the ratio of blood parasites to the number of cells of the selected cell type is calculated as NUM1 / NUM3.

[0112] like Figure 9 This study uses AI recognition algorithms to identify blood parasite samples. The AI ​​algorithm identifies normal red blood cells within a selected area S1 of the image, obtaining the total number of normal red blood cells NUM2 in the selected image. The AI ​​algorithm identifies normal red blood cells in the sample based on a dataset of normal red blood cell characteristics. The ratio of blood parasites to normal red blood cells is NUM1 / NUM2. The samples can be either mammalian or non-mammal.

[0113] like Figure 11A blood parasite detection device is disclosed for detecting blood parasites. The blood is pre-treated to obtain a test sample, which is then placed into a receiving cavity of a test sample receiving device. The device includes a control module, a camera assembly, a support assembly, and an image AI computing module. The control module is electrically connected to the camera assembly and the image AI computing module. The support assembly is used to hold the test sample receiving device. The image AI computing module includes a storage unit for storing AI feature data of the blood parasites. The camera assembly is used to capture images of the test sample to obtain test images. The image AI computing module is used to analyze the test images.

[0114] like Figure 11 The image AI computing module is located within the control module.

[0115] like Figure 13 The control module includes a network component; the image AI computing module is located in the server, and the network component is electrically connected to the image AI computing module via the network.

[0116] like Figure 14 The camera assembly includes a Z-axis slide assembly and a microscope camera assembly. The microscope camera assembly is mechanically connected to the slide in the Z-axis slide assembly, and the microscope camera assembly can move in the Z-axis direction. The drive module is electrically connected to the camera assembly and the Z-axis slide assembly, and the drive module is used to drive the microscope camera assembly to move in the Z-axis direction.

[0117] like Figures 12 to 14 It also includes a drive module, a control module that is electrically connected to the drive module, and a drive module that is electrically connected to the support component. The support component includes an X-slide assembly and a sample receiving device that is placed on the slide of the X-slide assembly. The drive module is used to drive the X-slide assembly to move in the X-axis direction, thereby moving the sample receiving device in the X-axis direction.

[0118] like Figure 14 The support components include a Y-slide assembly and an X-slide assembly placed on the slide of the Y-slide assembly. The drive module is used to drive the Y-slide assembly to move in the Y-axis direction, thereby moving the sample receiving device in the X-axis or Y-axis direction.

[0119] like Figure 15 The blood parasite detection device also includes circuit board A and circuit board B. The control module is located on circuit board A, and the drive module is located on circuit board B.

[0120] like Figure 16 The blood parasite detection device also includes a circuit board C, and the drive module and control module are located on the circuit board C.

[0121] like Figures 11 to 16The supporting components include a light source assembly, which includes a white light source. The white light emitted by the white light source is used to illuminate the receiving cavity.

[0122] like Figures 11 to 16 The supporting components include a light source assembly, which includes a purple light source. The purple light emitted by the purple light source is used to illuminate the receiving cavity.

[0123] like Figures 11 to 16 In some embodiments, the camera component is positioned above the support component. In other embodiments, the camera component is positioned below the support component.

[0124] like Figures 17 to 20 A sample container is used for blood parasite detection. The blood is pretreated to obtain a test sample, which is then added to the container chambers of the sample container, including container chamber A, container chamber B, and a sample inlet. The sample inlet is used to add the test sample. The sample inlet is connected to container chamber A and container chamber B. The liquid level of container chamber A is higher than that of container chamber B. Container chamber A is used to hold the test sample of species A, and container chamber B is used to hold the test sample of species B.

[0125] like Figure 17 The receiving chambers A and B are connected in series and are used to sequentially pour the liquid to be tested into receiving chambers A and B.

[0126] like Figure 18 The receiving chambers A and B are connected in parallel and are used to simultaneously fill the receiving chambers A and B with the liquid to be tested.

[0127] like Figure 19 and Figure 20 The two accommodating chambers A and B are connected in parallel but not interconnected, and are used to pour the liquid to be tested into accommodating chambers A and B respectively.

[0128] like Figures 17 to 20 The sample container also includes an exhaust port; the exhaust port is connected to the container cavity A; the exhaust port is connected to the container cavity B; and the exhaust port is connected to the outside atmosphere.

[0129] like Figures 17 to 20 The upper parts of cavity A and cavity B are made of transparent material.

[0130] In some embodiments, the lower part of the receiving cavity A and the lower part of the receiving cavity B are made of transparent material.

[0131] In some embodiments, receiving cavity A and receiving cavity B have the same bottom height.

[0132] In some embodiments, the receiving cavity A and the receiving cavity B have the same height at the top.

[0133] While the present invention has been described and illustrated with reference to preferred embodiments and several alternatives, the invention is not limited to the specific descriptions herein. Other alternatives or equivalent components may also be used to practice the invention.

Claims

1. A method for training an AI for detecting blood parasites, characterized in that, include: Collect blood from animals infected with parasites; Blood pretreatment is performed to obtain test samples; blood pretreatment includes diluting the blood with a diluent; Allow blood parasites to exist in their natural state in the liquid; The test sample is photographed to obtain test sample images; blood parasites in the test sample images are identified and labeled to obtain labeled images; AI is trained using the labeled images to obtain a blood parasite AI feature dataset; The AI ​​feature dataset of blood parasites, combined with corresponding AI algorithms, has the ability to identify blood parasites.

2. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, Illuminate the test sample with white light, the intensity of which is required to clearly show the image of the material inside the cell nucleus, and then take a picture of the test sample to obtain an image of the test sample; The blood parasite AI feature dataset, combined with the corresponding AI algorithm, has the ability to identify blood parasites in images of blood parasites collected under white light irradiation of the test sample.

3. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, The test sample is illuminated with ultraviolet light, and an image of the test sample is obtained. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to identify blood parasites in the blood parasite images collected from the test sample illuminated with ultraviolet light.

4. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, Includes any one of the following technical features: TF10: The parasite is either Hemoglobinozoa or Hepatobloc; TF20: The animal is a reptile, and the parasite is a blood fluke that lives inside red blood cells; TF30: The animal is a bird, and the parasite is a blood fluke that lives inside red blood cells; TF40: The animal is an amphibian, and the parasite is a blood fluke that lives inside red blood cells; TF50: The animal is a mammal, and the parasite is a blood fluke that lives inside red blood cells; TF60: The animal is a mammal, and the parasite is a liver tuft, which parasitizes leukocytes.

5. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, Includes any one of the following technical features: TK10: The infected animals are reptiles, which are infected by injecting the blood of animals infected with Blood Fissula. TK20: The infected animals are birds, which are infected by injecting blood from animals infected with Blood Fissula. TK30: The infected animals are amphibians, which are infected by injecting the blood of animals infected with Blood Cluster Parasite.

6. The AI ​​training method for blood parasite detection according to claim 5, Its features are, This includes: collecting blood samples from different stages of infection after the experimental animals are infected, wherein the blood parasite feature dataset includes a parasite infection stage feature dataset; the infection stage is expressed in days or hours.

7. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, Includes any one of the following technical features: TN10: The blood pretreatment includes staining of the test sample, which involves adding a staining agent to the blood; TN20: The blood pretreatment includes staining of the test sample, with the staining agent dissolved in the diluent in advance.

8. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, Normal red blood cells in the test sample images are identified and labeled to obtain labeled images. AI is trained using the labeled images to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset, combined with the corresponding AI algorithm, has the ability to identify normal red blood cells.

9. The AI ​​training method for blood parasite detection according to claim 1, characterized in that, White blood cells in the test sample images are identified and labeled to obtain labeled images. AI is trained using the labeled images to obtain a blood parasite AI feature dataset. The blood parasite AI feature dataset, combined with the corresponding AI algorithm, has the ability to identify white blood cells.

10. A method for detecting blood parasites, characterized in that, Blood pretreatment is performed to obtain test samples; blood pretreatment includes diluting the blood with a diluent; Allow blood parasites to exist in a natural state within the liquid; The sample is photographed to obtain an image of the sample. By combining the blood parasite AI feature dataset with the corresponding AI algorithm, the system identifies and detects parasites in the sample images. The blood parasite AI feature dataset was obtained by AI training on labeled blood parasite images.

11. The method for detecting blood parasites according to claim 10, characterized in that, The test sample is illuminated with white light, and an image of the test sample is taken. The intensity of the white light needs to be sufficient to clearly show the image of the substances inside the cell nucleus, thus obtaining an image of the test sample. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to have the ability to identify blood parasites in the blood parasite images collected by illuminating the test sample with white light.

12. The method for detecting blood parasites according to claim 10, characterized in that, The test sample is illuminated with ultraviolet light, and an image of the test sample is obtained. The blood parasite AI feature dataset is combined with the corresponding AI algorithm to identify blood parasites in the blood parasite images collected after the test sample is illuminated with ultraviolet light.

13. The method for detecting blood parasites according to claim 10, characterized in that, The AI ​​recognition algorithm is used to identify the test sample image, identify blood parasites in the sample within the selected area S1 of the image, and obtain the total number NUM1 of blood parasites in the sample of the selected image.

14. The method for detecting blood parasites according to claim 13, characterized in that, The volume of the sample corresponding to the selected area S1 is V1, and the volume content of blood parasites in the sample is NUM1 / V1.

15. The method for detecting blood parasites according to claim 10, characterized in that, The blood parasite AI feature dataset was obtained by training images labeled with manual annotations of the parasite infection stages; or by training blood images obtained from experimental animals after N days of infection.

16. The method for detecting blood parasites according to claim 13, characterized in that, Includes any one of the following technical features: TH10: Outputs the number of parasites at different stages of infection; TH20: Output the total number of selected parasites; TH30: Outputs the percentage of parasites at different infection stages; The blood parasite feature dataset was obtained by training on labeled blood parasite images.

17. The method for detecting blood parasites according to claim 10, characterized in that, Includes any one of the following technical features: TG10: The parasite is either Hemoglobinozoa or Hepatobloc; TG20: The animal is a reptile, and the parasite is a blood fluke that lives inside red blood cells; TG30: The animal is a bird, and the parasite is a blood fluke that lives inside red blood cells; TG40: The animal is an amphibian, and the parasite is a blood fluke that parasitizes red blood cells; TG50: The animal is a mammal, and the parasite is a blood fluke that lives inside red blood cells; TG60: The animal is a mammal, and the parasite is a liver tuft, which parasitizes leukocytes.

18. The method for detecting blood parasites according to claim 10, characterized in that, The AI ​​recognition algorithm is used to identify the detected sample image, identify the cells of the selected category in the sample within the selected area S1 of the image, and obtain the total number of cells of the selected category NUM3 in the sample of the selected image; the AI ​​recognition algorithm identifies the cells of the selected category in the sample based on the cell feature dataset of the selected category; The selected category of cell feature dataset is obtained by training labeled cells of the selected category.

19. The method for detecting blood parasites according to claim 18, characterized in that, Includes any one of the following technical features: TQ10: The selected cell category includes any one or more of the following: coagulation cells, coagulation cell clusters, leukocytes, immature nucleated cells, reticulocytes, and shadow cells; TQ20: The selected cell category content per unit volume of the sample is NUM3 / V1; TQ30: Calculate the ratio of blood parasites to the number of cells of the selected category = NUM1 / NUM3.

20. The method for detecting blood parasites according to claim 13, characterized in that, The AI ​​recognition algorithm is used to identify the detected sample image, identify normal red blood cells in the sample within the selected area S1 of the image, and obtain the total number of normal red blood cells NUM2 in the sample of the selected image; the AI ​​recognition algorithm identifies normal red blood cells in the sample based on the normal red blood cell feature dataset in the sample; the ratio of blood parasites to normal red blood cells = NUM1 / NUM2.

21. A blood parasite detection device for detecting blood parasites, wherein the blood is pretreated to obtain a test sample, and the test sample is added to the receiving cavity of the test sample receiving device, characterized in that, This includes a control module, camera components, support components, and an image AI computing module; The control module is electrically connected to the camera assembly; The control module is electrically connected to the image AI computing module; The support components are used to hold the sample container. The image AI computing module includes a storage unit, which is used to store AI feature data of blood parasites; The camera component is used to capture images of the test samples to obtain test images; The image AI computing module is used to analyze the detected image.

22. The blood parasite detection device according to claim 21, characterized in that, Includes one or more of the following technical features: TA10: The image AI calculation module is located within the control module; TA20: The control module includes a network component; the image AI computing module is located in the server, and the network component is electrically connected to the image AI computing module via a network. TA30: The camera assembly includes a Z-axis slide assembly and a microscope camera assembly. The microscope camera assembly is mechanically connected to the slide in the Z-axis slide assembly, and the microscope camera assembly can move in the Z-axis direction. TA40: It also includes a drive module, a control module and a drive module connected by electrical signals, and a drive module and a support component connected by electrical signals. The support component includes an X-slide assembly and a sample receiving device placed on the slide of the X-slide assembly. The drive module is used to drive the X-slide assembly to move in the X-axis direction, thereby moving the sample receiving device in the X-axis direction.

23. The blood parasite detection device according to claim 22, characterized in that, In TA40, the supporting components include Y The X-axis slide assembly and the X-axis slide assembly are placed on the slide of the Y-axis slide assembly. The drive module is used to drive the Y-axis slide assembly along the Y-axis. The direction of movement causes the sample container to move in the X-axis or Y-axis direction.

24. The blood parasite detection device according to claim 22, characterized in that, Includes any one of the following technical features: TB10: also includes circuit board A and circuit board B, the control module is disposed on circuit board A and the drive module is disposed on circuit board B; TB20: It also includes a circuit board C, on which the drive module and the control module are mounted. TC10: The support assembly includes a light source assembly, which includes a white light source. The white light emitted by the white light source is used to illuminate the receiving cavity. TC20: The support assembly includes a light source assembly, which includes a purple light source. The purple light emitted by the purple light source is used to illuminate the receiving cavity. TC30: The camera assembly is located above or below the support assembly.

25. A sample receiving device for detecting blood parasites, wherein the blood is pretreated to obtain a sample for testing, and the sample is added to the receiving cavity of the sample receiving device, characterized in that: Includes cavity A, cavity B, and sample dispensing port; The sample dispensing port is used to add the sample to be tested; The sample inlet is connected to cavity A; the sample inlet is connected to cavity B. The liquid holding height of cavity A is higher than that of cavity B; The cavity A is used to hold the test sample of species A; The containment chamber B is used to contain the test sample of species B.

26. The sample container according to claim 25, characterized in that, Includes one or more of the following technical features: TD10: The receiving cavity A and receiving cavity B are connected in series and are used to sequentially pour the liquid to be tested into receiving cavity A and receiving cavity B; TD20: The receiving cavity A and receiving cavity B are connected in parallel for the simultaneous filling of the liquid to be tested into receiving cavity A and receiving cavity B; TD30: The receiving cavity A and receiving cavity B are connected in parallel but not interconnected, and are used to pour the liquid to be tested into receiving cavity A and receiving cavity B respectively; TD40: also includes an exhaust port; the exhaust port is connected to the receiving cavity A; the exhaust port is connected to the receiving cavity B; the exhaust port is connected to the external atmosphere; TE10: The upper part of the receiving cavity A and the upper part of the receiving cavity B are made of transparent material; TE20: The lower parts of cavity A and cavity B are made of transparent material; TE30, wherein the receiving cavity A and the receiving cavity B have the same bottom height; TE40, the receiving cavity A and the receiving cavity B have the same height at the top.

Images