A magnetically controlled hollow organ cell collection device and storage medium

Abstract

This invention discloses a method, storage medium, and device for acquiring cells from hollow organs with magnetic control functionality, relating to the field of image processing technology. The method includes acquiring mechanical and displacement data of a magnet at a target location; obtaining motion characterization values ​​of the magnet based on the displacement data, and deriving multiple warning values ​​based on the force and motion characterization values; obtaining the upper limit of the current of the magnet controller, and obtaining real-time speed prompts for the magnet controller based on the force and motion characterization values; extracting features of multiple preset attributes from the mucosa at the target location, obtaining feature quantization values ​​corresponding to the multiple preset attributes; obtaining a recognition result based on a classification model and the feature quantization values ​​corresponding to the multiple preset attributes; and obtaining the structure at the target location based on a generative adversarial network model, combined with the recognition result, a standard hollow organ structure, and the feature quantization values ​​corresponding to the multiple preset attributes. This invention can effectively achieve the acquisition of hollow organs.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a method, storage medium, and device for acquiring cells from hollow organs with magnetic control function. Background Technology

[0002] Hollow organs, including the stomach, esophagus, intestines, and bladder, are difficult to detect through external imaging due to their indirect contact with the body surface and the presence of air interference. This poses a significant challenge to the early diagnosis and treatment of diseases, especially tumors. Currently, endoscopy is a commonly used method for detecting and diagnosing hollow organs, but its high cost and poor patient acceptance limit its widespread use. Accurate assessment of the mucosal health of hollow organs before endoscopic examination and screening of high-risk patients would greatly improve the cost-effectiveness of endoscopy and promote early diagnosis and treatment of diseases of hollow organs.

[0003] Currently, esophageal cell collection devices are available on the market, typically made of a pull-string, capsule, and sponge. After the patient swallows the capsule, the sponge is removed by pulling the string, scraping out esophageal mucosal cells for subsequent analysis. However, this type of cell collection device is only suitable for esophageal mucosa with a regular and narrow tubular structure, and is not suitable for other larger, hollow organs without a point of leverage. Furthermore, the way these devices scrape off cells is random and difficult to homogenize, leading to a series of problems in subsequent sample analysis, such as poor sample qualification rate and unstable results. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method, storage medium, and device for collecting cells from hollow organs with magnetic control function, which can effectively realize the collection of cells from hollow organs.

[0005] To achieve the above objectives, this invention provides a method for collecting cells from hollow organs with magnetic control function, specifically including the following steps:

[0006] Acquire the mechanical and displacement data of the magnet at the target position, and obtain the force characterization value of the contact surface of the magnet based on the mechanical data;

[0007] Based on displacement data, the motion characterization values ​​of the magnet are obtained, and multiple warning values ​​are obtained based on the force characterization values ​​and motion characterization values.

[0008] Based on the obtained multiple warning values, the upper limit value of the current of the magnet controller is obtained, and the real-time speed prompt information of the magnet controller is obtained according to the force characterization value and motion characterization value.

[0009] Based on the force characterization value, displacement data and motion characterization value, feature extraction of multiple preset attributes is performed on the mucosa at the target location to obtain the feature quantization value corresponding to the multiple preset attributes;

[0010] Based on the classification model and the feature quantization values ​​corresponding to multiple preset attributes, the recognition result is obtained;

[0011] Based on the adversarial generative network model, and combined with the recognition results, standard hollow organ structure, and feature quantization values ​​corresponding to multiple preset attributes, the structure of the target location is obtained.

[0012] Based on the above technical solutions,

[0013] The mechanical data refers to the reaction force of the digester's inner wall on the magnet;

[0014] The displacement data represents the actual trajectory of the magnet.

[0015] Based on the above technical solutions,

[0016] The force characteristics are the normal force on the magnet surface and the tangential force parallel to the magnet surface;

[0017] The motion characterization values ​​include the normal displacement of the magnet along a direction perpendicular to the magnet surface, and the tangential velocity of the magnet;

[0018] The tangential velocity is obtained from the derivative of the magnet's trajectory with respect to time.

[0019] Based on the above technical solutions,

[0020] The warning values ​​include warning values ​​for the magnitude of magnetic control force and warning values ​​for the extent of mucosal damage.

[0021] Based on the radial force and radial displacement of the magnet at the target position, when the mucosa stretching reaches the damage threshold, the magnetic control force value at this time is the warning value of the magnitude of the magnetic control force, and the area of ​​the stretched mucosa is the warning value of the range of mucosal damage.

[0022] Based on the above technical solutions,

[0023] The upper limit of the current of the magnet controller can be obtained by using the warning value of the magnitude of the magnetic control force;

[0024] The real-time speed indication information of the magnet controller is obtained by measuring the tangential velocity of the magnet.

[0025] Based on the above technical solutions,

[0026] The preset attributes include mucosal peristalsis frequency attribute, mucosal peristalsis height attribute, mucosal peristalsis position attribute, and mucosal curvature attribute;

[0027] The steps of extracting features from the mucosa at the target location using multiple preset attributes to obtain feature quantization values ​​corresponding to the multiple preset attributes include:

[0028] The ratio of the difference between the maximum and minimum radial force of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal peristalsis frequency characteristic.

[0029] The ratio of the difference between the location of the maximum radial force of the magnet and the location of the minimum radial force at the target location within a preset time period to the time difference is determined as the quantification value of the mucosal peristalsis height characteristic.

[0030] The ratio of the location to the time difference of the average radial force of the magnet at the target location within a preset time period is determined as the quantitative value of the mucosal peristalsis location feature.

[0031] The ratio of the curvature of the actual trajectory of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal curvature characteristic.

[0032] Based on the above technical solution, the feature quantification values ​​include mucosal peristalsis frequency feature quantification values, mucosal peristalsis height feature quantification values, mucosal peristalsis location feature quantification values, and mucosal curvature feature quantification values.

[0033] Based on the above technical solutions,

[0034] The classification model includes a feature fitting subnetwork and a classification subnetwork;

[0035] The process of obtaining the recognition result based on a classification model and according to the feature quantization values ​​corresponding to multiple preset attributes includes the following specific steps:

[0036] A feature fitting sub-network is used to fit the feature quantization values ​​corresponding to multiple preset attributes to obtain the determination coefficients.

[0037] Based on the obtained determination coefficients, a classification subnetwork is used for analysis to obtain the recognition results;

[0038] The identification results include the esophagus, cardia, gastric body, and antrum identified during cell collection in the upper digestive tract; the rectum, sigmoid colon, and transverse colon identified during cell collection in the lower digestive tract; and the locations identified during cell collection in other hollow organs.

[0039] The present invention provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for collecting cells from hollow organs with magnetic control function.

[0040] This invention provides a magnetically controlled hollow organ cell collection device, comprising:

[0041] The acquisition module is used to acquire the mechanical and displacement data of the magnet at the target position, and to acquire the force characterization value of the contact surface of the magnet based on the mechanical data.

[0042] The determination module is used to obtain the motion characterization value of the magnet based on the displacement data, and to obtain multiple warning values ​​based on the force characterization value and the motion characterization value;

[0043] The prompting module is used to obtain the upper limit value of the current of the magnet controller based on multiple warning values, and to obtain the real-time speed prompt information of the magnet controller according to the force characterization value and motion characterization value.

[0044] The quantization module is used to extract features of multiple preset attributes of the mucosa at the target location based on force characterization values, displacement data, and motion characterization values, and obtain feature quantization values ​​corresponding to multiple preset attributes.

[0045] The classification module is used to obtain the recognition result based on the classification model and the feature quantization values ​​corresponding to multiple preset attributes;

[0046] The execution module is used to obtain the structure of the target location based on the adversarial generative network model, combined with the recognition results, standard hollow organ structure and feature quantization values ​​corresponding to multiple preset attributes.

[0047] Compared with existing technologies, the advantages of this invention are as follows: By acquiring and based on the mechanical and displacement data of the magnet at the target position, the force characterization value and motion characterization value of the magnet are obtained. Then, based on the force characterization value and motion characterization value, multiple warning values ​​and real-time speed prompts from the magnet controller are obtained. Next, the upper limit value of the current of the magnet controller is obtained through the warning values. Then, features of multiple preset attributes of the mucosa at the target position are extracted to obtain feature quantification values. Finally, using a classification model and a generative adversarial network model, and based on the feature quantification values ​​of multiple preset attributes and the standard hollow organ structure, the structure at the target position is obtained. When collecting cells using magnetic control technology, this invention monitors and prompts based on the moving position, force, and direction of the magnet, allowing cell collection without damaging the organ mucosa. This provides a new method for identifying the location of hollow organs in patients and constructing the true structure of hollow organs in patients without endoscopic observation. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of a method for collecting cells from hollow organs with magnetic control function, as described in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments.

[0051] See Figure 1 As shown in the figure, an embodiment of the present invention provides a method for collecting cells from hollow organs with magnetic control function, which specifically includes the following steps:

[0052] S1: Obtain the mechanical and displacement data of the magnet at the target position, and obtain the force characterization value of the contact surface of the magnet based on the mechanical data;

[0053] In this invention, the mechanical data is the reaction force of the digester's inner wall on the magnet; the displacement data is the actual trajectory of the magnet.

[0054] In practical applications, a cell collector is constructed based on a magnet, and a force sensor and a displacement sensor are installed within the magnet. Once the cell collector is positioned at the target location (a hollow organ), the force sensor acquires the force acting on the magnet, and the displacement sensor acquires the magnet's displacement. The force sensor receives the force acting on the magnet, and the displacement sensor receives the magnet's displacement information. In this invention, the magnet is shaped like a low-profile cylinder, similar to a button battery.

[0055] In this invention, the force characteristics are the normal force on the magnet surface and the tangential force parallel to the magnet surface.

[0056] The mechanical data of the magnet at the target location includes its own force, the magnetic control force exerted on the magnet by external magnetic devices, and the force exerted by the digestive tract wall. In practical applications, the force sensor can receive and transmit the attractive force of the external magnetic devices on the magnet at the target location, as well as the reaction force of the digestive tract wall on the magnet at the target location.

[0057] It should be noted that the force decomposition mentioned below refers to the decomposition of the reaction force exerted by the digestive tract wall on the magnet at the target location. The force decomposition yields the force characterization value of the contact surface of the magnet at the target location. This force characterization value can be the normal force acting on the surface of the magnet at the target location and the tangential force parallel to the surface of the magnet at the target location.

[0058] The force can be projected along a direction perpendicular to the surface of the magnet at the target position to obtain the normal force; the force can be projected along a direction parallel to the surface of the magnet at the target position to obtain the tangential force.

[0059] S2: Based on the displacement data, obtain the motion characterization value of the magnet, and based on the force characterization value and motion characterization value, obtain multiple warning values;

[0060] In this invention, the motion characterization values ​​include the normal displacement of the magnet along a direction perpendicular to the magnet surface, and the tangential velocity of the magnet; the tangential velocity is obtained based on the derivative of the magnet's motion trajectory with time.

[0061] The displacement data is the actual trajectory of the magnet at the target position, which is the relationship between the coordinate position and time. The motion characteristics of the magnet at the target position can be the normal displacement and the tangential velocity of the magnet at the target position.

[0062] Specifically, the normal displacement of the magnet can be obtained by projecting the actual trajectory of the magnet at the target position along a direction perpendicular to the magnet surface; the velocity of the magnet can be obtained by calculating the derivative of the actual trajectory of the magnet at the target position with time, and then the velocity can be projected along a direction parallel to the magnet surface to obtain the tangential velocity of the magnet.

[0063] In this invention, the warning values ​​include a warning value for the magnitude of the magnetic control force and a warning value for the range of mucosal damage. Based on the radial force and radial displacement of the magnet at the target position, when the mucosal stretching reaches the damage threshold, the magnetic control force value at this time is the warning value for the magnitude of the magnetic control force, and the area of ​​the stretched mucosa is the warning value for the range of mucosal damage.

[0064] As the attraction of the external magnetic device to the magnet at the target location increases, the mucosa of the digestive tract will be stretched. When the stretching exceeds a certain degree, the mucosal surface may rupture, bleed, or even perforate.

[0065] Specifically, by combining the radial force, radial displacement, and elastic modulus of the target magnet, the mucosal stretching value is calculated. When the mucosal stretching reaches the damage threshold, the magnetic control force value at this time is the warning value for the magnitude of the magnetic control force, and the area of ​​the stretched mucosa at this time is the warning value for the range of mucosal damage.

[0066] S3: Based on the obtained multiple warning values, obtain the upper limit value of the current of the magnet controller, and obtain the real-time speed prompt information of the magnet controller according to the force characterization value and motion characterization value;

[0067] In this invention, the upper limit of the current of the magnet controller is obtained by the warning value of the magnitude of the magnetic control force; and the real-time speed prompt information of the magnet controller is obtained by the tangential velocity of the magnet.

[0068] The magnetic control force increases with the increase of the control current. When the tangential force on the magnet at the target position exceeds a certain threshold, it indicates that the inner wall of the digestive tract at the position of the magnet changes from a plane to a curved surface.

[0069] Specifically, the upper limit of the magnet controller current can be calculated from the warning value of the magnetic control force; the real-time tangential velocity of the magnet at the target position provides a prompt on the movement speed of the external magnet, such as "Speed ​​too fast, please slow down." The magnet controller is located externally and is used to control the magnetism of external magnetic devices.

[0070] S4: Based on the force characterization value, displacement data and motion characterization value, perform feature extraction on the mucosa at the target location using multiple preset attributes to obtain the feature quantization value corresponding to the multiple preset attributes;

[0071] In this invention, the preset attributes include mucosal peristalsis frequency attribute, mucosal peristalsis height attribute, mucosal peristalsis position attribute, and mucosal curvature attribute.

[0072] Specifically, feature extraction methods can be used to extract mucosal features and obtain quantified feature values. These methods can be manual feature extraction combined with image feature analysis algorithms, such as pixel neighborhood mean calculation and maximum pixel value extraction, to calculate the quantified feature values. Alternatively, deep learning feature extraction methods, such as Convolutional Neural Networks (CNN) and UNet++, can be used. The specific method can be selected based on the features of the preset attributes, and no limitation is imposed here. In this invention, feature extraction of the mucosa is performed using force representation values, displacement, and motion representation values ​​to obtain corresponding quantified feature values. This achieves the quantification calculation of features of various preset attributes of the mucosa, making the quantified feature values ​​more comprehensive and rich. This allows for accurate and intuitive site identification and image transfer based on these multiple quantified feature values.

[0073] In this invention, multiple preset attributes are extracted from the mucosa at the target location to obtain feature quantization values ​​corresponding to the multiple preset attributes. Specific steps include:

[0074] The ratio of the difference between the maximum and minimum radial force of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal peristalsis frequency characteristic.

[0075] The ratio of the difference between the location of the maximum radial force of the magnet and the location of the minimum radial force at the target location within a preset time period to the time difference is determined as the quantification value of the mucosal peristalsis height characteristic.

[0076] The ratio of the location to the time difference of the average radial force of the magnet at the target location within a preset time period is determined as the quantitative value of the mucosal peristalsis location feature.

[0077] The ratio of the curvature of the actual trajectory of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal curvature characteristic.

[0078] In this invention, the feature quantification values ​​include mucosal peristalsis frequency feature quantification values, mucosal peristalsis height feature quantification values, mucosal peristalsis location feature quantification values, and mucosal curvature feature quantification values.

[0079] S5: Based on the classification model and according to the feature quantization values ​​corresponding to multiple preset attributes, the recognition result is obtained;

[0080] In this invention, the classification model includes a feature fitting subnetwork and a classification subnetwork.

[0081] In this invention, based on a classification model and according to the feature quantization values ​​corresponding to multiple preset attributes, the recognition result is obtained. Specific steps include:

[0082] S501: The feature fitting sub-network is used to fit the feature quantization values ​​corresponding to multiple preset attributes to obtain the determination coefficients;

[0083] S502: Based on the obtained decision coefficients, a classification subnetwork is used for analysis to obtain the recognition results;

[0084] The identification results include the esophagus, cardia, gastric body, and antrum identified during cell collection in the upper digestive tract; the rectum, sigmoid colon, and transverse colon identified during cell collection in the lower digestive tract; and the locations identified during cell collection in other hollow organs.

[0085] That is, the classification model can be a trained machine learning classifier, which can learn a machine learning algorithm model with classification capabilities through samples. The machine learning classifier in this invention is used to classify different sets of feature values ​​into one class of normal or abnormal results. Specifically, it can be a classifier that uses at least one machine learning model for classification. The machine learning model can be one or more of the following: neural networks (e.g., convolutional neural networks, backpropagation neural networks, etc.), logistic regression models, support vector machines, decision trees, random forests, perceptrons, and other machine learning models.

[0086] As part of the training of the machine learning model, the training input is the quantized values ​​of various features, such as the quantized values ​​of mucosal peristalsis frequency, mucosal peristalsis height, mucosal peristalsis location, and mucosal curvature. Through training, a classifier is established to establish the correspondence between the set of feature values ​​and whether there is an abnormality in the part to be identified, so that the preset classifier has the ability to judge whether the classification result corresponding to the part to be identified is a normal result or an abnormal result.

[0087] In this invention, the classifier is a multi-classifier. For example, when collecting cells in the upper digestive tract, the sites are the esophagus, cardia, gastric body, antrum, etc.; when collecting cells in the lower digestive tract, they are classified into the rectum, sigmoid colon, transverse colon, etc.

[0088] This invention fully considers the impact of quantified feature values ​​of multiple different attributes on the accuracy and intuitiveness of digestive tract site identification. By extracting features with richer information and quantifying and comprehensively processing features of multiple different attributes, the rationality of feature value quantification is improved. Compared with traditional methods that only consider single feature information and single statistical comparison methods, this invention provides a new approach to digestive tract site identification when there is no image observation.

[0089] S6: Based on the adversarial generative network model, and combined with the recognition results, standard hollow organ structure and feature quantization values ​​corresponding to multiple preset attributes, the structure of the target location is obtained.

[0090] Specifically, multiple feature quantification values, recognition results, and standard digestive tract images are input into a trained generative adversarial network (GAN) model. For example, a GAN network model is selected. In one specific implementation, the quantification values ​​of mucosal peristalsis frequency, mucosal peristalsis height, mucosal peristalsis location, mucosal curvature, and site recognition results, along with the standard digestive tract image, are used as input to the trained GAN model. The output of the GAN model is a real image of the patient's stomach or esophagus / colon. Understandably, the real images of the patient's stomach, esophagus, and colon provide a reference for subsequent pathological diagnosis by physicians.

[0091] In one possible implementation, the force exerted on the inner wall of the digestive tract is... The force is decomposed to obtain the normal force perpendicular to the surface of the magnet at the target position. and tangential force parallel to the surface of the magnet at the target position .

[0092] In one possible implementation, the actual trajectory of the magnet at the target location. ,Will The normal displacement perpendicular to the magnet surface at the target position is obtained by decomposition. and tangential displacement parallel to the surface of the magnet at the target position , The tangential velocity of the magnet is obtained by differentiating with respect to time. .

[0093] In one possible implementation, the specific steps for determining the warning values ​​for the magnitude of the magnetic control force and the range of mucosal damage are as follows:

[0094] a: The radius of the magnet contact surface at the target location is The distance a magnet travels from its initial contact with the mucosal surface to the point where it is attracted by magnetic force, causing localized stretching of the digestive tract wall, is... At this time, the normal force is ;

[0095] b: Determine the local stretching amount of the mucosa: ,when At this time, the corresponding This is the warning value for the magnitude of magnetic control force. ;

[0096] c: The length of the mucosa after stretching is The warning value for the extent of mucosal damage is... .

[0097] In one possible implementation, a warning value for the magnitude of the magnetic control force is obtained. Then, the upper limit value of the current of the external magnet controller can be obtained; through the force characterization value and speed characterization value, the real-time speed prompt information of the magnet controller can be obtained.

[0098] Specifically, it is known that, under constant conditions such as magnetic field strength, conductor material, and conductor length, the magnetic force is positively correlated with the current. The upper limit of the current for the external magnet controller is... ,in These are characterization coefficients related to the magnetic field and conductor involved in this invention; based on the force characterization value and velocity characterization value ,when When the resistance encountered during lateral movement is significant, a message will be displayed stating, "There may be a bend in the digestive tract wall here; please operate with caution." At that time, the message "Speed ​​too fast, please slow down" will be displayed. When the current level is controlled, the message "Horizontal movement has stopped. Please pay attention to controlling the current level. Cell collection can proceed" will be displayed.

[0099] In one possible implementation, when there is mucosal peristalsis on the digestive tract wall, the radial force on the contact surface of the magnet at the target location... Peristalsis changes over time; attributes such as mucosal peristalsis frequency, height, and location can all be obtained through... Characterize the changes over time;

[0100] Specifically, Time to Within a given time period, the quantified value of the mucosal peristalsis frequency characteristic can be represented by the difference between the maximum and minimum radial force values ​​over that period. ; Time to Within a given time period, the quantified value of the mucosal peristalsis height characteristic can be represented by the difference between the location of the maximum radial force and the location of the minimum radial force over a certain time period. ; Time to Within a given time period, the quantitative value of the mucosal peristaltic positional characteristics can be represented by the location of the average radial force over a certain time period. .

[0101] When the mucosa undergoes a large curvature change, the curvature of the actual trajectory of the magnet at the target position will also change significantly. Therefore, the quantification value of the mucosa curvature feature can be characterized by the curvature of the actual trajectory of the magnet at the target position.

[0102] Specifically, Time to The actual trajectory of the magnet at the target position within a given time is an arc. The quantified value of mucosal curvature is then... .

[0103] By analyzing the relationship between the magnitude and position of the radial force of the magnet at the target location and the change over time, as well as the analysis of the actual motion trajectory of the magnet at the target location, the quantification results of the mucosal peristalsis frequency attribute, mucosal peristalsis height attribute, mucosal peristalsis position attribute, and mucosal curvature attribute are obtained. This is not only simple to calculate, but also helps to improve the accuracy of subsequent site identification and image correction.

[0104] In this invention, the trained machine learning classifier includes a feature fitting sub-network and a classification sub-network. The step of inputting the quantified values ​​of multiple attribute features of the digestive tract mucosa into the trained machine learning classifier for classification to obtain the site identification classification result includes: using the quantified values ​​of mucosal peristalsis frequency features, mucosal peristalsis height features, mucosal peristalsis location features, and mucosal curvature features of the feature fitting sub-network for fitting processing to obtain a determination coefficient; and using the classification sub-network for analysis based on the determination coefficient to obtain the identification result.

[0105] Specifically, a feature fitting subnetwork is used to fit the quantified values ​​of multiple attribute features of the digestive tract mucosa. Based on the fitting results, the corresponding weights for fitting the quantified values ​​of multiple attribute features of each digestive tract mucosa are determined. The quantified values ​​of the mucosal peristalsis frequency feature from the above embodiment are then used as the basis for further fitting. Quantitative values ​​of mucosal peristalsis height characteristics Quantitative values ​​of mucosal peristalsis location characteristics Quantitative values ​​of mucosal curvature characteristics For example, decision trees and random forests are used to determine... , , , The corresponding weights are respectively Then the fused feature value is:

[0106] .

[0107] In this invention, by fusing and calculating the quantitative values ​​of multiple attribute features of the digestive tract mucosa, the information features of the digestive tract mucosa are made richer and the quantification is more accurate, which is conducive to improving the efficiency and accuracy of site identification.

[0108] In one possible implementation, the mucosal peristalsis frequency characteristic is quantized. Quantitative values ​​of mucosal peristalsis height characteristics Quantitative values ​​of mucosal peristalsis location characteristics Quantitative values ​​of mucosal curvature characteristics and body part recognition results The pre-trained adversarial neural network model inputs standard digestive tract images to perform image transfer on standard digestive tract structures, thereby obtaining the patient's actual stomach, esophagus, or colorectal structure.

[0109] The method for collecting cells from hollow organs with magnetic control functionality according to this invention first acquires and obtains the force and motion characteristics of the magnet based on the mechanical and displacement data of the magnet at the target location. Then, based on the force and motion characteristics, multiple warning values ​​and real-time speed prompts from the external magnet controller are obtained. Next, the upper limit of the current of the external magnet controller is obtained through the warning values. Then, features of multiple preset attributes are extracted from the mucosa at the target location to obtain feature quantification values. Finally, a classification model and a generative adversarial network model are used, and based on the feature quantification values ​​of the multiple preset attributes and the standard digestive tract structure, the structure at the target location is obtained. When collecting cells using magnetic control technology, this invention monitors and prompts based on the magnet's movement position, force, and direction of movement to obtain digestive tract site identification and the patient's actual digestive tract structure.

[0110] In one possible implementation, the present invention also provides a non-transitory computer-readable storage medium located in a PLC (Programmable Logic Controller) controller. The storage medium stores a computer program that, when executed by a processor, implements the steps of the automotive diagnostic interface information security testing method described below:

[0111] Acquire the mechanical and displacement data of the magnet at the target position, and obtain the force characterization value of the contact surface of the magnet based on the mechanical data;

[0112] Based on displacement data, the motion characterization values ​​of the magnet are obtained, and multiple warning values ​​are obtained based on the force characterization values ​​and motion characterization values.

[0113] Based on the obtained multiple warning values, the upper limit value of the current of the magnet controller is obtained, and the real-time speed prompt information of the magnet controller is obtained according to the force characterization value and motion characterization value.

[0114] Based on the force characterization value, displacement data and motion characterization value, feature extraction of multiple preset attributes is performed on the mucosa at the target location to obtain the feature quantization value corresponding to the multiple preset attributes;

[0115] Based on the classification model and the feature quantization values ​​corresponding to multiple preset attributes, the recognition result is obtained;

[0116] Based on the adversarial generative network model, and combined with the recognition results, standard hollow organ structure, and feature quantization values ​​corresponding to multiple preset attributes, the structure of the target location is obtained.

[0117] Storage media may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. Computer-readable storage media may be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in connection with an instruction execution system, apparatus, or device.

[0118] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wireline, optical fiber, RF, etc., or any suitable combination thereof.

[0119] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0120] The present invention provides a hollow organ cell collection device with magnetic control function, comprising an acquisition module, a determination module, a prompting module, a quantification module, a classification module and an execution module.

[0121] In this invention, the acquisition module acquires the mechanical and displacement data of the magnet at the target location, and obtains the force characterization value of the contact surface of the magnet based on the mechanical data; the determination module acquires the motion characterization value of the magnet based on the displacement data, and obtains multiple warning values ​​based on the force characterization value and motion characterization value; the prompting module obtains the upper limit value of the current of the magnet controller based on the obtained multiple warning values, and obtains the real-time speed prompt information of the magnet controller based on the force characterization value and motion characterization value; the quantization module extracts features of multiple preset attributes of the mucosa at the target location based on the force characterization value, displacement data, and motion characterization value, and obtains the feature quantization value corresponding to the multiple preset attributes; the classification module obtains the recognition result based on the classification model and the feature quantization value corresponding to the multiple preset attributes; and the execution module obtains the structure at the target location based on the generative adversarial network model, combined with the recognition result, standard hollow organ structure, and the feature quantization value corresponding to the multiple preset attributes.

[0122] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

[0123] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

Claims

1. A device for collecting cells from hollow organs with magnetic control function, characterized in that, include: The acquisition module is used to acquire the mechanical and displacement data of the magnet at the target position, and to acquire the force characterization value of the contact surface of the magnet based on the mechanical data. The determination module is used to obtain the motion characterization value of the magnet based on the displacement data, and to obtain multiple warning values ​​based on the force characterization value and the motion characterization value; The prompting module is used to obtain the upper limit value of the current of the magnet controller based on multiple warning values, and to obtain the real-time speed prompt information of the magnet controller according to the force characterization value and motion characterization value. The quantization module is used to extract features of multiple preset attributes of the mucosa at the target location based on force characterization values, displacement data, and motion characterization values, and obtain feature quantization values ​​corresponding to multiple preset attributes. The classification module is used to obtain the recognition result based on the classification model and the feature quantization values ​​corresponding to multiple preset attributes; The execution module is used to obtain the structure of the target location based on the adversarial generative network model, combined with the recognition results, standard hollow organ structure and feature quantization values ​​corresponding to multiple preset attributes; The warning values ​​include a warning value for the magnitude of the magnetic control force and a warning value for the extent of mucosal damage; Based on the radial force and radial displacement of the magnet at the target position, when the mucosa stretching reaches the damage threshold, the magnetic control force value at this time is the warning value of the magnitude of the magnetic control force, and the area of ​​the stretched mucosa is the warning value of the range of mucosal damage. Among them, the upper limit of the current of the magnet controller can be obtained by using the warning value of the magnitude of the magnetic control force; The real-time speed indication information of the magnet controller is obtained by measuring the tangential velocity of the magnet; Specifically, the determination of the warning values ​​for the magnitude of magnetic control force and the range of mucosal damage includes: Obtain the radius of the magnet contact surface at the target location, the distance the magnet moves from the moment it contacts the mucosal surface until it is attracted by the magnetic force and causes local stretching of the digestive tract wall, and the normal force at this time; Based on the normal force at this time and the radius of the magnet contact surface at the target position, the local stretching amount of the mucosa is calculated and determined. When the calculated local stretching amount of the mucosa is greater than or equal to the threshold, the calculated local stretching amount of the mucosa is the warning value of the magnitude of the magnetic control force. Based on the radius of the magnet contact surface at the target location, the distance the magnet moves from the moment it contacts the mucosal surface to the point when it is attracted by the magnetic force and causes local stretching of the digestive tract wall, and the amount of local stretching of the mucosa, the length of the mucosa after stretching is calculated, and the warning value of the mucosal damage range is calculated based on the length of the mucosa after stretching. Among them, the upper limit of the current of the magnet controller is determined by the product between the warning value of the magnetic control force and the characterization coefficient; in, The mechanical data refers to the reaction force of the digester's inner wall on the magnet; The displacement data represents the actual trajectory of the magnet. in, The force characteristics are the normal force on the magnet surface and the tangential force parallel to the magnet surface; The motion characterization values ​​include the normal displacement of the magnet along a direction perpendicular to the magnet surface, and the tangential velocity of the magnet; The tangential velocity is obtained from the derivative of the magnet's trajectory with respect to time; in, The preset attributes include mucosal peristalsis frequency attribute, mucosal peristalsis height attribute, mucosal peristalsis position attribute, and mucosal curvature attribute; The steps of extracting features from the mucosa at the target location using multiple preset attributes to obtain feature quantization values ​​corresponding to the multiple preset attributes include: The ratio of the difference between the maximum and minimum radial force of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal peristalsis frequency characteristic. The ratio of the difference between the location of the maximum radial force of the magnet and the location of the minimum radial force at the target location within a preset time period to the time difference is determined as the quantification value of the mucosal peristalsis height characteristic. The ratio of the location to the time difference of the average radial force of the magnet at the target location within a preset time period is determined as the quantitative value of the mucosal peristalsis location feature. The ratio of the curvature of the actual trajectory of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal curvature characteristic. The feature quantization values ​​include mucosal peristalsis frequency feature quantization values, mucosal peristalsis height feature quantization values, mucosal peristalsis location feature quantization values, and mucosal curvature feature quantization values; in, The classification model includes a feature fitting subnetwork and a classification subnetwork; The process of obtaining the recognition result based on a classification model and according to the feature quantization values ​​corresponding to multiple preset attributes includes the following specific steps: A feature fitting sub-network is used to fit the feature quantization values ​​corresponding to multiple preset attributes to obtain the determination coefficients. Based on the obtained determination coefficients, a classification subnetwork is used for analysis to obtain the recognition results; The identification results include the esophagus, cardia, gastric body, and antrum identified during cell collection in the upper digestive tract; the rectum, sigmoid colon, and transverse colon identified during cell collection in the lower digestive tract; and the locations identified during cell collection in other hollow organs.

2. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of a method for collecting cells from hollow organs with magnetic control functionality. The method for collecting cells from hollow organs with magnetic control function specifically includes the following steps: Acquire the mechanical and displacement data of the magnet at the target position, and obtain the force characterization value of the contact surface of the magnet based on the mechanical data; Based on displacement data, the motion characterization values ​​of the magnet are obtained, and multiple warning values ​​are obtained based on the force characterization values ​​and motion characterization values. Based on the obtained multiple warning values, the upper limit value of the current of the magnet controller is obtained, and the real-time speed prompt information of the magnet controller is obtained according to the force characterization value and motion characterization value. Based on the force characterization value, displacement data and motion characterization value, feature extraction of multiple preset attributes is performed on the mucosa at the target location to obtain the feature quantization value corresponding to the multiple preset attributes; Based on the classification model and the feature quantization values ​​corresponding to multiple preset attributes, the recognition result is obtained; Based on the adversarial generative network model, and combined with the recognition results, standard hollow organ structure and feature quantization values ​​corresponding to multiple preset attributes, the structure of the target location is obtained; The warning values ​​include a warning value for the magnitude of the magnetic control force and a warning value for the extent of mucosal damage; Based on the radial force and radial displacement of the magnet at the target position, when the mucosa stretching reaches the damage threshold, the magnetic control force value at this time is the warning value of the magnitude of the magnetic control force, and the area of ​​the stretched mucosa is the warning value of the range of mucosal damage. Among them, the upper limit of the current of the magnet controller can be obtained by using the warning value of the magnitude of the magnetic control force; The real-time speed indication information of the magnet controller is obtained by measuring the tangential velocity of the magnet; Specifically, the determination of the warning values ​​for the magnitude of magnetic control force and the range of mucosal damage includes: Obtain the radius of the magnet contact surface at the target location, the distance the magnet moves from the moment it contacts the mucosal surface until it is attracted by the magnetic force and causes local stretching of the digestive tract wall, and the normal force at this time; Based on the normal force at this time and the radius of the magnet contact surface at the target position, the local stretching amount of the mucosa is calculated and determined. When the calculated local stretching amount of the mucosa is greater than or equal to the threshold, the calculated local stretching amount of the mucosa is the warning value of the magnitude of the magnetic control force. Based on the radius of the magnet contact surface at the target location, the distance the magnet moves from the moment it contacts the mucosal surface to the point when it is attracted by the magnetic force and causes local stretching of the digestive tract wall, and the amount of local stretching of the mucosa, the length of the mucosa after stretching is calculated, and the warning value of the mucosal damage range is calculated based on the length of the mucosa after stretching. Among them, the upper limit of the current of the magnet controller is determined by the product between the warning value of the magnetic control force and the characterization coefficient; in, The mechanical data refers to the reaction force of the digester's inner wall on the magnet; The displacement data represents the actual trajectory of the magnet. in, The force characteristics are the normal force on the magnet surface and the tangential force parallel to the magnet surface; The motion characterization values ​​include the normal displacement of the magnet along a direction perpendicular to the magnet surface, and the tangential velocity of the magnet; The tangential velocity is obtained from the derivative of the magnet's trajectory with respect to time; in, The preset attributes include mucosal peristalsis frequency attribute, mucosal peristalsis height attribute, mucosal peristalsis position attribute, and mucosal curvature attribute; The steps of extracting features from the mucosa at the target location using multiple preset attributes to obtain feature quantization values ​​corresponding to the multiple preset attributes include: The ratio of the difference between the maximum and minimum radial force of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal peristalsis frequency characteristic. The ratio of the difference between the location of the maximum radial force of the magnet and the location of the minimum radial force at the target location within a preset time period to the time difference is determined as the quantification value of the mucosal peristalsis height characteristic. The ratio of the location to the time difference of the average radial force of the magnet at the target location within a preset time period is determined as the quantitative value of the mucosal peristalsis location feature. The ratio of the curvature of the actual trajectory of the magnet at the target location within a preset time period to the time difference is determined as the quantified value of the mucosal curvature characteristic. The feature quantization values ​​include mucosal peristalsis frequency feature quantization values, mucosal peristalsis height feature quantization values, mucosal peristalsis location feature quantization values, and mucosal curvature feature quantization values; in, The classification model includes a feature fitting subnetwork and a classification subnetwork; The process of obtaining the recognition result based on a classification model and according to the feature quantization values ​​corresponding to multiple preset attributes includes the following specific steps: A feature fitting sub-network is used to fit the feature quantization values ​​corresponding to multiple preset attributes to obtain the determination coefficients. Based on the obtained determination coefficients, a classification subnetwork is used for analysis to obtain the recognition results; The identification results include the esophagus, cardia, gastric body, and antrum identified during cell collection in the upper digestive tract; the rectum, sigmoid colon, and transverse colon identified during cell collection in the lower digestive tract; and the locations identified during cell collection in other hollow organs.

Images