Cell invasion and metastasis capability monitoring method and system based on radio frequency image analysis

Abstract

The embodiment of the invention provides a cell invasion and metastasis capability monitoring method and system based on radio frequency image analysis, and belongs to the technical field of artificial intelligence. The method comprises the following steps: continuously capturing echo signals of a lower chamber by using a millimeter wave technology to obtain a first echo signal corresponding to a first moment and a second echo signal corresponding to a second moment of the lower chamber; acquiring first image information corresponding to the first moment and second image information corresponding to the second moment of the lower room by using an image sensor; performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell; performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell; fusing the first feature parameter and the second feature parameter to obtain a target feature parameter corresponding to the target cell; and monitoring the invasion and metastasis capability of the target cells according to the target characteristic parameters to obtain a target detection result corresponding to the target cells.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and system for monitoring cell invasion and metastasis ability based on radio frequency image analysis. Background Art

[0002] In patients with pancytopenia, measuring the migration of specific cells across the vascular wall is a key step in measuring the immune response. The study of such cell motility is widely used, especially in the invasion and metastasis of tumor cells. In addition, the study of the migration ability of oncoblasts in the blood is one of the methods to measure the ability of tumor metastasis. Tumor metastasis is one of the biological characteristics of malignant tumors. At certain stages of the tumor metastasis process, tumor cells show strong motility, such as tumor cells separating from the primary tumor, invading adjacent tissues, and then passing through the vascular wall to enter the blood circulation and passing through the vascular wall to enter the secondary site. Therefore, the study of cell motility plays an important role in clarifying the mechanism of tumor cell invasion and metastasis.

[0003] The research method of tumor motility is mainly to detect the ability of tumor cells to actively migrate in vitro, generally including scratch experiments, induced metastasis experiments, etc. In induced metastasis, the upper chamber is a serum-free cell suspension, and the lower chamber is a serum-containing culture medium. Certain components in the serum, such as some cytokines, are essential for cell growth. Therefore, cells have a tendency to these factors. This tendency of cells is used to promote the migration of cells to the lower chamber. After migration, the cells will adhere to the bottom of the polycarbonate membrane. The number of cells at the bottom of the polycarbonate membrane is used to show the cell migration ability. Different cells will show different movement and migration abilities here. However, when the existing technology uses image analysis algorithms to detect tumor movement and determine the growth of cells in the lower chamber, there are problems with insufficient detection accuracy and reliability.

[0004] Therefore, it is urgent to design a technical solution to solve at least one of the above technical problems. Summary of the invention

[0005] The main purpose of the embodiments of the present invention is to provide a method and system for monitoring cell invasion and metastasis ability based on radio frequency image analysis, aiming to solve the problem of insufficient detection accuracy and reliability when using image analysis algorithms to detect tumor movement and determine the growth of lower chamber cells in related technologies.

[0006] In a first aspect, an embodiment of the present invention provides a method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis, comprising:

[0007] In the induced transfer, the millimeter wave technology is used to continuously capture the echo signal of the lower chamber to obtain a first echo signal corresponding to the first moment and a second echo signal corresponding to the second moment of the lower chamber;

[0008] Using an image sensor to collect first image information corresponding to the lower chamber at the first moment and second image information corresponding to the second moment, the lower chamber being a culture medium containing serum;

[0009] Performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell;

[0010] Performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell;

[0011] fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell;

[0012] The invasion and metastasis capabilities of the target cells are monitored according to the target characteristic parameters to obtain target detection results corresponding to the target cells.

[0013] In a second aspect, an embodiment of the present invention provides a cell invasion and metastasis ability monitoring system based on radio frequency imaging analysis, comprising:

[0014] A signal acquisition module, used to continuously capture the echo signal of the lower chamber using millimeter wave technology during induced transfer, and obtain a first echo signal corresponding to the lower chamber at a first moment and a second echo signal corresponding to the second moment;

[0015] An image acquisition module, used for acquiring, by using an image sensor, first image information corresponding to the first moment and second image information corresponding to the second moment of the lower chamber, wherein the lower chamber is a culture medium containing serum;

[0016] A first analysis module, configured to perform signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell;

[0017] A second analysis module, configured to perform image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell;

[0018] A parameter fusion module, used for fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell;

[0019] The result determination module is used to monitor the invasion and metastasis ability of the target cells according to the target characteristic parameters to obtain the target detection results corresponding to the target cells.

[0020] In a third aspect, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the methods for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided in the specification of the present invention are implemented.

[0021] In a fourth aspect, an embodiment of the present invention further provides a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided in the specification of the present invention.

[0022] The embodiment of the present invention provides a method and system for monitoring cell invasion and metastasis ability based on radio frequency image analysis, the method comprising: in induced metastasis, using millimeter wave technology to continuously capture echo signals of the lower chamber, obtaining a first echo signal corresponding to the lower chamber at the first moment and a second echo signal corresponding to the second moment, so that the changes of the cells can be monitored in real time, which is helpful to timely discover the metastasis of the cells and provide support for obtaining accurate first characteristic parameters in the subsequent process; using an image sensor to collect the first image information corresponding to the lower chamber at the first moment and the second image information corresponding to the second moment; performing signal change analysis on the first echo signal and the second echo signal to obtain the first characteristic parameter corresponding to the target cell; performing image change analysis on the first image information and the second image information to obtain the second characteristic parameter corresponding to the target cell; fusing the first characteristic parameter and the second characteristic parameter to obtain the target characteristic parameter corresponding to the target cell; monitoring the invasion and metastasis ability of the target cell according to the target characteristic parameter, and obtaining the target detection result corresponding to the target cell. By combining millimeter wave technology and image sensor data, the method can more comprehensively capture the dynamic characteristics of the target cell, thereby improving the detection accuracy of the cell invasion and metastasis ability. Furthermore, this method combines the first characteristic parameter with the second characteristic parameter to provide richer information. Combining signal analysis with image analysis can comprehensively reflect cell behavior, reduce errors that may be caused by a single data source, and effectively reduce errors in the experimental process, improve the reliability of experimental results, and thus significantly improve the monitoring effect of the invasion and metastasis ability of target cells, providing important support for tumor research and clinical applications. It also solves the problem of insufficient detection accuracy and reliability when using image analysis algorithms to detect tumor movement and determine the growth of lower chamber cells in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 A schematic diagram of a process of monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of the module structure of a cell invasion and metastasis ability monitoring system based on radio frequency image analysis provided by an embodiment of the present invention;

[0026] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] The embodiment of the present invention provides a method and system for monitoring cell invasion and metastasis based on radio frequency imaging analysis. The method for monitoring cell invasion and metastasis based on radio frequency imaging analysis can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0031] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0032] Please refer to Figure 1 , Figure 1 A schematic flow chart of a method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided in an embodiment of the present invention.

[0033] like Figure 1 As shown, the method for monitoring cell invasion and metastasis ability based on radio frequency image analysis includes steps S101 to S106.

[0034] Step S101: During induced transfer, the millimeter wave technology is used to continuously capture echo signals of the lower chamber to obtain a first echo signal of the lower chamber corresponding to a first moment and a second echo signal corresponding to a second moment.

[0035] For example, the research method of tumor motility is mainly to detect the ability of tumor cells to actively migrate in vitro, generally including scratch test, induced metastasis test, etc. In induced metastasis, the upper chamber is a serum-free cell suspension, and the lower chamber is a serum-containing culture medium. Certain components in the serum, such as some cytokines, are essential for cell growth, so cells have a tendency to these factors. This tendency of cells is used to promote the migration of cells to the lower chamber. After migration, the cells will adhere to the bottom of the polycarbonate membrane. The number of cells at the bottom of the polycarbonate membrane shows the cell migration ability. Different cells will show different movement and migration abilities here.

[0036] Exemplarily, in the lower chamber, changes in the number of cells will affect the echo signal when the millimeter wave technology is continuously capturing the echo signal of the lower chamber. Therefore, the millimeter wave technology is used to continuously capture the echo signal of the lower chamber, and the first echo signal of the lower chamber at the first moment and the second echo signal at the second moment are obtained, and the first moment and the second moment are different moments.

[0037] For example, a millimeter-wave radar sensor is used to place the lower chamber to be studied (such as a culture dish containing cells) within the detectable range of the millimeter-wave radar. The material of the lower chamber needs to be able to allow millimeter waves to penetrate or reflect to obtain effective echo signals. The millimeter-wave radar sensor is then aimed at the lower chamber to ensure that the radar beam can cover the entire lower chamber area in order to capture accurate echo signals, and then the millimeter-wave radar system is started to start emitting millimeter waves and receiving echo signals from the lower chamber, and the first echo signal captured by the millimeter-wave radar at the first moment, and the second echo signal captured by the millimeter-wave radar at the second moment after a certain time interval.

[0038] Step S102: using an image sensor to collect first image information corresponding to the lower chamber at the first moment and second image information corresponding to the second moment, wherein the lower chamber is a culture medium containing serum.

[0039] For example, a suitable image sensor, such as CCD, CMOS, etc., is selected according to the research needs, taking into account performance indicators such as resolution, sensitivity, and noise level. An image sensor that can work within the required spectral range is selected, especially one that can have a good imaging effect on the observed cell or tissue features.

[0040] Exemplarily, the image sensor is installed in a suitable position to ensure that a clear image of the entire lower room or a specific area can be obtained. Note that interference or noise is prevented during the image capture process. Thus, at the first moment, the millimeter-wave radar is used to capture the first echo signal while the image sensor is also used to collect the first image information corresponding to the lower room. And at the second moment, the millimeter-wave radar is used to capture the second echo signal while the image sensor is also used to collect the second image information corresponding to the lower room.

[0041] Step S103: performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell.

[0042] Exemplarily, the first echo signal and the second echo signal are compared in the time domain to obtain the amplitude change, duration of the amplitude change, and change in waveform characteristics between the first echo signal and the second echo signal, and then the statistical values ​​such as the mean, maximum value, minimum value, and standard deviation of the first echo signal and the second echo signal are calculated to evaluate the first change information between the first echo signal and the second echo signal in the time domain. The first echo signal and the second echo signal are then Fourier transformed respectively, converted to the frequency domain, and the intensity and phase changes of each frequency component are analyzed, and the enhancement or attenuation of the corresponding frequency within the frequency range related to the target cell characteristics and the phase information of the frequency range are determined, so as to determine the second change information between the first echo signal and the second echo signal in the frequency domain according to the enhancement or attenuation of the frequency and the phase information of the frequency range.

[0043] Exemplarily, the moving speed of the target cell is determined by the rate of signal change in the first change information and the second change information, and the moving direction of the target cell is determined by the change in signal strength and the signal phase difference in the first change information and the second change information. Thus, the moving speed and moving direction of the target cell are determined as the corresponding first characteristic parameters.

[0044] In some embodiments, the signal change analysis of the first echo signal and the second echo signal to obtain the first characteristic parameter corresponding to the target cell includes: determining a target window and a target step size, and cutting the first echo signal according to the target window and the target step size to obtain multiple first window signals; cutting the second echo signal according to the target window and the target step size to obtain multiple second window signals; one-to-one correspondence between the multiple first window signals and the multiple second window signals to obtain multiple target window groups, each target window group includes a first window signal and a second window signal corresponding to the first window signal; respectively serializing the first window signal and the second window signal in the target window group to obtain first sequence information corresponding to the first window signal and second sequence information corresponding to the second window signal; determining the information change entropy corresponding to the target window group according to the first sequence information and the second sequence information; performing curve fitting according to the information change entropy to obtain a target curve, and determining the first characteristic parameter corresponding to the target cell according to the target curve.

[0045] Exemplarily, the size of the target window is defined. The target window is used to capture a time segment of the echo signal. The choice of the target window size depends on the frequency content of the signal, the sampling rate, and the dynamic change rate of the target cell. A shorter target window is suitable for capturing rapid changes, and a longer target window is suitable for observing slower changes. The target step size refers to the interval between each movement of the target window when cutting the echo signal. The step size can be the same as the size of the target window (no overlap) or smaller than the size of the target window (with overlap). A smaller target step size can provide higher time domain resolution, but will increase the amount of calculation. Select an appropriate target step size according to the experimental requirements to ensure that the continuous changes of the signal can be captured.

[0046] Exemplarily, the first echo signal is cut using the determined target window and target step size, thereby dividing the entire first echo signal into a plurality of first window signals. Each first window signal is a segment in the first echo signal, representing the change of the signal in a specific time period. Then, the second echo signal is subjected to a similar cutting operation using the same target window and target step size to obtain a plurality of second window signals. The second window signal corresponds to the first window signal of the first echo signal in time.

[0047] Exemplarily, a plurality of first window signals are matched one-to-one with a plurality of second window signals to form a plurality of target window groups. Each target window group includes a first window signal and a second window signal corresponding thereto. It is ensured that the corresponding relationship is based on the same position.

[0048] For example, the first echo signal is cut into first window signals W1_1, W1_2, and W1_3, and the second echo signal is cut into second window signals W2_1, W2_2, and W2_3, then the target window group can be expressed as ((W1_1, W2_1), (W1_2, W2_2), (W1_3, W2_3).

[0049] Exemplarily, the first window signal in each target window group is serialized and converted into a first sequence information. Serialization is to arrange the data points in the time domain signal in a certain order to form a sequence. This process is helpful for subsequent statistical analysis and information entropy calculation. Similarly, the second window signal in each target window group is serialized to obtain the corresponding second sequence information.

[0050] Exemplarily, the information change entropy is calculated based on the first sequence information and the second sequence information in each target window group. Information change entropy is an indicator to measure the uncertainty or complexity of the signal. The entropy value can be calculated by comparing the difference or similarity between the two sequences. For example, the relative entropy (KL divergence) is used to measure the distribution difference of the two sequences. For example, if the first sequence information and the second sequence information show a large difference in the time domain or frequency domain, the information change entropy will be higher, indicating that the signal change is more drastic or the uncertainty is larger.

[0051] Exemplarily, the calculated information change entropy is subjected to curve fitting. The purpose of curve fitting is to find a mathematical function that can best fit these entropy value changes. Common fitting methods such as polynomial fitting, exponential fitting, or nonlinear fitting can be used, depending on the trend of entropy value changes. The curve obtained by fitting is called the target curve. The target curve should be able to smoothly describe the changing trend of information change entropy over time or window position, thereby revealing the overall law and trend of signal changes.

[0052] Exemplarily, a target curve is used to indicate the movement of target cells in different regions of the lower chamber between a first moment and a second moment. A curve mean corresponding to the target curve is obtained, and a first movement intensity corresponding to the target cell is determined using the curve mean; and a first movement direction corresponding to the target cell is determined using the maximum value of the target curve. Thus, a first characteristic parameter corresponding to the target cell is determined according to the first movement intensity and the first movement direction.

[0053] In some embodiments, determining the first characteristic parameter corresponding to the target cell according to the target curve includes: obtaining a maximum value corresponding to the target curve, and determining a first moving direction corresponding to the target cell according to the maximum value; integrating the target curve according to the target window and the target step size to obtain a target area corresponding to the target window; obtaining a first cell content corresponding to a unit entropy, and determining the number of cells corresponding to the target cell under the target window according to the target area and the first cell content; determining a first cell density corresponding to the target window according to the cell number; and determining the first characteristic parameter corresponding to the target cell according to the first cell density and the first moving direction.

[0054] Exemplarily, a target reference position is determined, which is a reference point for determining the corresponding moving direction of the target cell, and then a maximum value is found from the fitted target curve. This maximum value usually represents the point where the signal changes most dramatically, which may correspond to the most significant activity or state change of the target cell at a specific time or under certain conditions. According to the position where the maximum value of the target curve appears and the target reference position, the first moving direction of the target cell is determined.

[0055] Exemplarily, the target window and target step size are used to perform an integration operation on the target curve. The result of the integration will obtain the target area corresponding to the target window. The integration operation can be simply understood as calculating the area of ​​the curve within the target window. The integration result of the target curve is used as the target area, and the area value reflects the cumulative change of the signal within the target window. A larger target area indicates that the signal has a larger change range within the window, which may mean that the number of target cells is larger or the activity is more intense.

[0056] Exemplarily, the cell content corresponding to the unit entropy is determined by experimental data or empirical formula. This unit entropy can be a fixed entropy value, reflecting the number of cells corresponding to the unit change. Therefore, according to the target area and the cell content corresponding to the unit entropy, the number of target cells under the target window is calculated, and then the calculated number of cells is used as the actual number of target cells under the target window. A larger number of cells may mean that the cell population density is larger or the cell activity is more active. According to the calculated number of cells divided by the area of ​​the target window, the first cell density under the target window is determined.

[0057] Exemplarily, the first cell density and the first moving direction are determined as the first characteristic parameters of the target cells. The first cell density reflects the spatial distribution characteristics of the target cells and the first moving direction reflects the dynamic behavior characteristics of the target cells.

[0058] Step S104: performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell.

[0059] Exemplarily, ensure that the first image information and the second image information are aligned in spatial position. Image registration technology, such as a registration method based on feature point matching or pixel intensity, can be used to align the two images into the same coordinate system. On the aligned image, calculate the difference between the first image information and the second image information to obtain a difference image.

[0060] Exemplarily, edges, corners, grayscale changes in image regions, etc. are extracted from the difference image, thereby performing change detection on the extracted features. It is determined which features have changed significantly between the first image information and the second image information. These change points may correspond to the location, morphological changes, or other important features of the target cells.

[0061] Exemplarily, the changed area in the second image information relative to the first image information is marked with colors and labels, and then image processing tools such as histogram analysis and regional statistics are used to calculate statistics such as the average grayscale value, variance, maximum value and minimum value of the changed area to evaluate the degree of change.

[0062] Exemplarily, a target reference position is determined, which is a reference point for determining the movement direction corresponding to the target cell. The point where the signal changes most dramatically is determined based on the maximum value corresponding to the change area, which may correspond to the most significant activity or state change of the target cell at a specific time or under certain conditions. The movement direction of the target cell is determined based on the position where the maximum value corresponding to the change area appears and the target reference position.

[0063] Exemplarily, the time difference between the first moment and the second moment is calculated, and the number of changes corresponding to the change area is counted, so as to determine the movement speed of the target cell according to the number of changes and the time difference, and thus the movement direction and movement speed are determined as the second characteristic parameters corresponding to the target cell.

[0064] In some embodiments, the image change analysis of the first image information and the second image information to obtain the second feature parameter corresponding to the target cell includes: using the coding layer of the edge segmentation model to perform feature extraction on the first image information to obtain the corresponding first coding feature; using the edge perception layer of the edge segmentation model to perform edge extraction on the first image information to obtain the corresponding first edge feature; using the feature fusion layer of the edge segmentation model to perform feature superposition on the first coding feature and the first edge feature to obtain a first enhanced feature; using the reordering layer of the edge segmentation model to perform feature fusion and channel rearrangement according to the first coding feature, the first edge feature and the first enhanced feature to obtain a first fused feature; using the adaptive channel attention layer of the edge segmentation model to determine the first target convolution kernel corresponding to the first fused feature, and performing feature learning on the first fused feature according to the first target convolution kernel to obtain a first target feature; using the edge recognition layer of the edge segmentation model to perform feature learning on the first fused feature according to the The first target feature is used to determine the first edge information corresponding to the first image information; the encoding layer is used to extract features from the second image information to obtain the corresponding second encoding feature; the edge perception layer is used to extract edges from the second image information to obtain the corresponding second edge feature; the feature fusion layer is used to superimpose the second encoding feature and the second edge feature to obtain a second enhanced feature; the reordering layer is used to perform feature fusion and channel rearrangement according to the second encoding feature, the second edge feature and the second enhanced feature to obtain a second fused feature; the adaptive channel attention layer is used to determine the second target convolution kernel corresponding to the second fusion feature, and feature learning is performed on the second fusion feature according to the second target convolution kernel to obtain the second target feature; the edge recognition layer is used to determine the second edge information corresponding to the second image information according to the second target feature; image change analysis is performed according to the first edge information and the second edge information to obtain the second feature parameter corresponding to the target cell.

[0065] Exemplarily, the first image information is processed using the coding layer of the edge segmentation model to extract high-level features and generate corresponding first coding features. The coding layer is usually used to capture the global information and low-level feature representation of the image, and then the edge perception layer is used to extract the edge of the first image information. The purpose of the edge perception layer is to identify important edge structures in the image and generate corresponding first edge features. The first coding feature and the first edge feature are then input into the feature fusion layer for superposition to generate a first enhanced feature. This process combines the global features extracted by the coding layer and the edge features extracted by the edge layer, thereby improving the expressiveness of the features.

[0066] Exemplarily, in the reordering layer, the first coding feature, the first edge feature and the first enhancement feature are fused and then the channels are rearranged, and one channel feature of each of the three features is taken in order and arranged in sequence to form a first fused feature. The purpose is to enable subsequent convolution operations to effectively capture the information interaction between different features without using a large convolution kernel.

[0067] Exemplarily, the first fused feature is input to the adaptive channel attention layer, which determines the first target convolution kernel corresponding to the first fused feature. The first fused feature is learned using the determined first target convolution kernel to obtain the first target feature. The adaptive channel attention mechanism can help the model focus on more important features. Thus, according to the first target feature, the edge recognition layer is used to determine the first edge information corresponding to the first image information. This process will help to accurately identify the cell edges in the image.

[0068] Similarly, the encoding layer is used to extract features of the second image information to generate the corresponding second encoding features, and then the second image information is input into the edge perception layer to extract the second edge features and identify the edge information in the image. Then, the second encoding features and the second edge features are superimposed using the feature fusion layer to generate the second enhanced features. Then, the second encoding features, the second edge features, and the second enhanced features are feature fused and the channels are rearranged using the reordering layer to obtain the second fused features. Then, the second fused features are input into the adaptive channel attention layer to determine the second target convolution kernel, and then the second fused features are learned based on the second target convolution kernel to finally obtain the second target features, and finally, the edge recognition layer is used to identify the second edge information of the second image information according to the second target features.

[0069] Exemplarily, first pixel information corresponding to the first edge information is obtained from the first image information and second pixel information corresponding to the second edge information is obtained from the second image information, so that the moving direction corresponding to the target cell is determined according to the difference between the first pixel information and the second pixel information.

[0070] In some embodiments, the performing image change analysis according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell includes: obtaining a first edge image corresponding to the first edge information and a first position information corresponding to the first edge information from the first image information, and obtaining a second edge image corresponding to the second edge information and a second position information corresponding to the second edge information from the second image information; performing pixel difference analysis on the first edge image to obtain a first significant area corresponding to the first edge image, and determining a first target position corresponding to the first significant area according to the first position information; performing pixel difference analysis on the second edge image to obtain a second significant area corresponding to the second edge image, and determining a second target position corresponding to the second significant area according to the second position information; determining a second moving direction corresponding to the target cell according to the first target position and the second target position; determining a second cell content corresponding to a preset pixel, and obtaining a first pixel distribution corresponding to the second edge image; determining a second cell density corresponding to the target cell according to the second cell content and the first pixel distribution; and determining the second characteristic parameter corresponding to the target cell according to the second cell density and the second moving direction.

[0071] Exemplarily, the first edge image is obtained by segmenting the first image information according to the first edge information, and the first edge image includes the first edge information and also includes the information included in the first edge information. The first position information corresponding to the first edge information is extracted from the first image information according to the first edge information, and the first position information is usually expressed as the center position of the first edge image. Similarly, the second edge image is obtained by segmenting the second image information according to the second edge information, and the second edge image includes the second edge information and also includes the information included in the second edge information. The second position information corresponding to the second edge information is extracted from the second image information according to the second edge information, and the second position information is usually expressed as the center position of the second edge image.

[0072] Exemplarily, the pixel difference calculation is performed on any two edge regions in the first edge image, and then the edge region corresponding to the larger pixel value in the two edge regions when the pixel difference is the largest is determined as the region of significant change in the first edge image, and the edge region is determined as the first significant region. The first target position corresponding to the first significant region is then obtained from the first position information. Similarly, the pixel difference calculation is performed on any two edge regions in the second edge image, and then the edge region corresponding to the larger pixel value in the two edge regions when the pixel difference is the largest is determined as the region of significant change in the second edge image, and the edge region is determined as the second significant region. The second target position corresponding to the second significant region is then obtained from the second position information.

[0073] Exemplarily, the second moving direction of the target cell is calculated according to the coordinate difference between the first target position and the second target position. The second moving direction can be expressed as a vector or an angle pointing from the first target position to the second target position.

[0074] Exemplarily, the second cell content corresponding to different pixels is obtained through experiments, and then the number of cells corresponding to the second edge image is obtained by multiplying the first pixel distribution in the second edge image by the second cell content, and then the second cell density corresponding to the second image information is determined by dividing the number of cells by the image area corresponding to the second image information.

[0075] Exemplarily, the second cell density and the second movement direction are determined as the second characteristic parameters corresponding to the target cells, that is, the second characteristic parameters can be expressed as a joint description of the cell density and the movement direction, which is used to comprehensively characterize the state of the target cells.

[0076] In some embodiments, the pixel difference analysis of the first edge image to obtain the first salient area corresponding to the first edge image includes: using the pixel segmentation layer of the salient information extraction model to perform region segmentation according to the pixel differences between the first edge images to obtain multiple first segmented images; using the information extraction layer of the salient information extraction model to perform multi-channel feature extraction on the first segmented image to obtain a first feature map; using the information fusion layer of the salient information extraction model to perform feature fusion on the first feature map and the first position information to obtain a first fusion map; using the data processing layer of the salient information extraction model to perform data analysis on the first fusion map to obtain a first mean and a first standard deviation; using the salient feature extraction layer of the salient information extraction model to perform feature extraction according to the first mean and the first standard deviation combined with the first fusion map to obtain a first salient feature; using the salient area determination layer of the salient information extraction model to perform saliency classification on the first segmented image according to the first salient feature to obtain the first salient area corresponding to the first edge image; wherein the first salient feature is obtained according to the following formula:

[0077]

[0078] Among them, X i represents the first significant feature corresponding to the i-th first segmented image, sigmoid represents the activation function, γ, θ and represents the training parameters corresponding to the salient feature extraction layer, α i represents the first mean corresponding to the i-th first segmented image, β i represents the first standard deviation corresponding to the i-th first segmented image, T irepresents the i-th first fusion graph.

[0079] Exemplarily, the pixel segmentation layer of the significant information extraction model performs difference analysis on the pixels in the first edge image to identify the pixel areas with significant changes. Pixel intensity difference or gradient difference can be used as a judgment criterion. Based on the analysis results of the pixel difference, the pixel segmentation layer of the significant information extraction model is used to perform regional segmentation on the first edge image to obtain multiple first segmented images. Then, the information extraction layer of the significant information extraction model is used to perform multi-channel feature extraction on each first segmented image. Multi-channel feature extraction can include the extraction of features such as color, texture, and shape. It can be achieved through structures such as convolutional layers and pooling layers in convolutional neural networks (CNN). The feature extraction results of each channel form a feature map, and the first feature map can be obtained by combining these feature maps.

[0080] Exemplarily, the information fusion layer of the salient information extraction model is used to perform feature fusion on the first feature map and the first position information. Feature fusion methods may include concatenation, weighted fusion, attention mechanism, etc. The fusion result forms a first fusion map. The data processing layer of the salient information extraction model is used to perform data analysis on the first fusion map. Data analysis may include statistical analysis (such as mean and standard deviation calculation) and data smoothing processing, etc. By calculating the mean and standard deviation of the first fusion map, the first mean and the first standard deviation are obtained respectively. Then, the salient feature extraction layer of the salient information extraction model is used to extract salient features based on the calculated first mean and first standard deviation in combination with the first fusion map, wherein the first salient feature is obtained according to the following formula:

[0081]

[0082] Among them, X i represents the first significant feature corresponding to the first segmented image of the i-th image, sigmoid represents the activation function, γ, θ and represents the training parameters corresponding to the salient feature extraction layer, α i represents the first mean corresponding to the i-th first segmented image, β i represents the first standard deviation corresponding to the i-th first segmented image, T i Represents the i-th first fusion graph.

[0083] Exemplarily, the salient region determination layer of the salient information extraction model is used to perform saliency classification on the first segmented image according to the extracted first salient features. Saliency classification can be achieved by a trained classifier (such as a support vector machine SVM, a random forest, a neural network, etc.). The classification result divides the region in the first segmented image into a salient region and a non-salient region, and finally obtains the first salient region corresponding to the first edge image.

[0084] Similarly, in some embodiments, pixel difference analysis is performed on the second edge image to obtain a second salient area corresponding to the second edge image, including: using a pixel segmentation layer of a salient information extraction model to perform area segmentation according to pixel differences between the second edge images to obtain a plurality of second segmented images; using an information extraction layer of the salient information extraction model to perform multi-channel feature extraction on the second segmented image to obtain a second feature map; using an information fusion layer of the salient information extraction model to perform feature fusion on the second feature map and the second position information to obtain a second fusion map; using a data processing layer of the salient information extraction model to perform data analysis on the second fusion map to obtain a second mean and a second standard deviation; using a salient feature extraction layer of the salient information extraction model to perform feature extraction based on the second mean and the second standard deviation combined with the second fusion map to obtain a second salient feature; using a salient area determination layer of the salient information extraction model to perform saliency classification on the second segmented image according to the second salient feature to obtain the second salient area corresponding to the second edge image.

[0085] The specific implementation manner is the same as that of performing pixel difference analysis on the first edge image to obtain the first salient area corresponding to the first edge image, and the present application will not elaborate on it.

[0086] Step S105: fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell.

[0087] Exemplarily, a machine learning model is used to perform parameter fusion on the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell.

[0088] In some embodiments, the fusing of the first characteristic parameter and the second characteristic parameter to obtain the target characteristic parameter corresponding to the target cell includes: determining a first weight corresponding to the first characteristic parameter and a second weight corresponding to the second characteristic parameter; performing information fusing of the first characteristic parameter and the second characteristic parameter according to the first weight and the second weight to obtain the target characteristic parameter corresponding to the target cell.

[0089] For example, based on expert experience or historical data, a first weight corresponding to a first characteristic parameter obtained after continuous echo signal capture of the lower chamber by millimeter wave technology and a second weight corresponding to a second characteristic parameter obtained after collecting the corresponding image of the lower chamber by an image sensor are obtained, and then the first characteristic parameter and the second characteristic parameter are weightedly fused using the first weight and the second weight to obtain the target characteristic parameter corresponding to the target cell. Before weighted fusion, the first characteristic parameter and the second characteristic parameter are converted to the same dimension to improve the reliability of the target characteristic parameter.

[0090] Step S106: monitoring the invasion and metastasis capabilities of the target cells according to the target characteristic parameters to obtain target detection results corresponding to the target cells.

[0091] Exemplarily, a preset threshold is determined, and the target characteristic parameter is compared with the preset threshold. When the target characteristic parameter exceeds the preset threshold, it is determined that the target detection result corresponding to the invasion and metastasis ability monitoring of the target cell is strong in metastasis ability; when the target characteristic parameter does not exceed the preset threshold, it is determined that the target detection result corresponding to the invasion and metastasis ability monitoring of the target cell is weak in metastasis ability.

[0092] See also Figure 2 , Figure 2 A cell invasion and metastasis ability monitoring system 200 based on radio frequency imaging analysis is provided in an embodiment of the present application. The cell invasion and metastasis ability monitoring system 200 based on radio frequency imaging analysis includes a signal acquisition module 201, an image acquisition module 202, a first analysis module 203, a second analysis module 204, a parameter fusion module 205, and a result determination module 206. The signal acquisition module 201 is used to continuously capture the echo signal of the lower chamber using millimeter wave technology during induction of metastasis, and obtain a first echo signal corresponding to the lower chamber at a first moment and a second echo signal corresponding to the second moment; the image acquisition module 202 is used to use an image sensor to collect the first image information corresponding to the lower chamber at the first moment and the second echo signal corresponding to the second moment The second image information corresponding to the second moment, the lower chamber is a culture medium containing serum; a first analysis module 203, used to perform signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell; a second analysis module 204, used to perform image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell; a parameter fusion module 205, used to fuse the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell; a result determination module 206, used to monitor the invasion and metastasis ability of the target cell according to the target characteristic parameter to obtain a target detection result corresponding to the target cell.

[0093] In some embodiments, the first analysis module 203, in the process of performing signal change analysis on the first echo signal and the second echo signal to obtain the first characteristic parameter corresponding to the target cell, performs:

[0094] Determining a target window and a target step size, and cutting the first echo signal according to the target window and the target step size to obtain a plurality of first window signals;

[0095] Cutting the second echo signal according to the target window and the target step size to obtain a plurality of second window signals;

[0096] Acquire multiple target window groups by making a one-to-one correspondence between a plurality of the first window signals and a plurality of the second window signals, each of the target window groups including a first window signal and a second window signal corresponding to the first window signal;

[0097] Sequencing the first window signal and the second window signal in the target window group respectively to obtain first sequence information corresponding to the first window signal and second sequence information corresponding to the second window signal;

[0098] Determine the information change entropy corresponding to the target window group according to the first sequence information and the second sequence information;

[0099] A target curve is obtained by performing curve fitting according to the information change entropy, and the first characteristic parameter corresponding to the target cell is determined according to the target curve.

[0100] In some embodiments, during the process of determining the first characteristic parameter corresponding to the target cell according to the target curve, the first analysis module 203 performs:

[0101] Obtaining a maximum value corresponding to the target curve, and determining a first moving direction corresponding to the target cell according to the maximum value;

[0102] Performing an integration operation on the target curve according to the target window and the target step length to obtain a target area corresponding to the target window;

[0103] Obtaining a first cell content corresponding to a unit entropy, and determining the number of cells corresponding to the target cells in the target window according to the target area and the first cell content;

[0104] Determine a first cell density corresponding to the target window according to the cell number;

[0105] The first characteristic parameter corresponding to the target cell is determined according to the first cell density and the first moving direction.

[0106] In some embodiments, the second analysis module 204 performs, during the process of performing image change analysis on the first image information and the second image information to obtain the second characteristic parameter corresponding to the target cell:

[0107] Using the coding layer of the edge segmentation model to perform feature extraction on the first image information to obtain a corresponding first coding feature;

[0108] Using the edge perception layer of the edge segmentation model to extract edges of the first image information, to obtain corresponding first edge features;

[0109] Using the feature fusion layer of the edge segmentation model, the first coding feature and the first edge feature are superimposed to obtain a first enhanced feature;

[0110] Using the reordering layer of the edge segmentation model, feature fusion and channel rearrangement are performed according to the first encoding feature, the first edge feature and the first enhanced feature to obtain a first fused feature;

[0111] Determine a first target convolution kernel corresponding to the first fusion feature by using the adaptive channel attention layer of the edge segmentation model, and perform feature learning on the first fusion feature according to the first target convolution kernel to obtain a first target feature;

[0112] Determine first edge information corresponding to the first image information according to the first target feature using the edge recognition layer of the edge segmentation model;

[0113] Using the coding layer to extract features from the second image information to obtain corresponding second coding features;

[0114] Using the edge perception layer to perform edge extraction on the second image information to obtain corresponding second edge features;

[0115] Using the feature fusion layer to perform feature superposition on the second coding feature and the second edge feature to obtain a second enhanced feature;

[0116] Using the reordering layer to perform feature fusion and channel rearrangement according to the second coding feature, the second edge feature and the second enhanced feature, to obtain a second fused feature;

[0117] Determine a second target convolution kernel corresponding to the second fusion feature by using the adaptive channel attention layer, and perform feature learning on the second fusion feature according to the second target convolution kernel to obtain a second target feature;

[0118] Determining second edge information corresponding to the second image information according to the second target feature using the edge recognition layer;

[0119] An image change analysis is performed according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell.

[0120] In some embodiments, the second analysis module 204, during the process of performing image change analysis according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell, performs:

[0121] Obtaining a first edge image corresponding to the first edge information and first position information corresponding to the first edge information from the first image information, and obtaining a second edge image corresponding to the second edge information and second position information corresponding to the second edge information from the second image information;

[0122] Performing pixel difference analysis on the first edge image to obtain a first salient area corresponding to the first edge image, and determining a first target position corresponding to the first salient area according to the first position information;

[0123] Performing pixel difference analysis on the second edge image to obtain a second salient area corresponding to the second edge image, and determining a second target position corresponding to the second salient area according to the second position information;

[0124] Determining a second moving direction corresponding to the target cell according to the first target position and the second target position;

[0125] determining a second cell content corresponding to a preset pixel, and obtaining a first pixel distribution corresponding to the second edge image;

[0126] Determining a second cell density corresponding to the target cells according to the second cell content and the first pixel distribution;

[0127] The second characteristic parameter corresponding to the target cell is determined according to the second cell density and the second moving direction.

[0128] In some implementations, the second analysis module 204, in the process of performing pixel difference analysis on the first edge image to obtain the first salient area corresponding to the first edge image, executes:

[0129] Using a pixel segmentation layer of a salient information extraction model to perform region segmentation according to pixel differences between the first edge images to obtain a plurality of first segmented images;

[0130] Using the information extraction layer of the salient information extraction model to perform multi-channel feature extraction on the first segmented image to obtain a first feature map;

[0131] Using the information fusion layer of the salient information extraction model to perform feature fusion on the first feature map and the first position information to obtain a first fusion map;

[0132] Using the data processing layer of the salient information extraction model to perform data analysis on the first fusion image to obtain a first mean and a first standard deviation;

[0133] Using the salient feature extraction layer of the salient information extraction model to extract features according to the first mean and the first standard deviation in combination with the first fusion graph, to obtain a first salient feature;

[0134] Using the salient region determination layer of the salient information extraction model to perform saliency classification on the first segmented image according to the first salient feature, to obtain the first salient region corresponding to the first edge image;

[0135] The first significant feature is obtained according to the following formula:

[0136]

[0137] Among them, X i represents the first significant feature corresponding to the i-th first segmented image, sigmoid represents the activation function, γ, θ and represents the training parameters corresponding to the salient feature extraction layer, α i represents the first mean corresponding to the i-th first segmented image, β i represents the first standard deviation corresponding to the i-th first segmented image, T i represents the i-th first fusion graph.

[0138] In some embodiments, the parameter fusion module 205 performs, during the process of fusing the first characteristic parameter and the second characteristic parameter to obtain the target characteristic parameter corresponding to the target cell:

[0139] Determine a first weight corresponding to the first characteristic parameter and a second weight corresponding to the second characteristic parameter;

[0140] The first characteristic parameter and the second characteristic parameter are subjected to information fusion according to the first weight and the second weight to obtain the target characteristic parameter corresponding to the target cell.

[0141] In some embodiments, the cell invasion and metastasis ability monitoring system 200 based on radio frequency image analysis can be applied to a terminal device.

[0142] It should be noted that, those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the cell invasion and metastasis ability monitoring system 200 based on radio frequency imaging analysis described above can refer to the corresponding process in the aforementioned embodiment of the cell invasion and metastasis ability monitoring method based on radio frequency imaging analysis, and will not be repeated here.

[0143] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention.

[0144] like Figure 3 As shown, the terminal device 300 includes a processor 301 and a memory 302, and the processor 301 and the memory 302 are connected via a bus 303, such as an I2C (Inter-integrated Circuit) bus.

[0145] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0146] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0147] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present invention, and does not constitute a limitation on the terminal device to which the embodiment of the present invention is applied. The specific server may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] The processor is used to run a computer program stored in the memory, and implement any one of the methods for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided by the embodiments of the present invention when executing the computer program.

[0149] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program:

[0150] In the induced transfer, the millimeter wave technology is used to continuously capture the echo signal of the lower chamber to obtain a first echo signal corresponding to the first moment and a second echo signal corresponding to the second moment of the lower chamber;

[0151] Using an image sensor to collect first image information corresponding to the lower chamber at the first moment and second image information corresponding to the second moment, the lower chamber being a culture medium containing serum;

[0152] Performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell;

[0153] Performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell;

[0154] fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell;

[0155] The invasion and metastasis capabilities of the target cells are monitored according to the target characteristic parameters to obtain target detection results corresponding to the target cells.

[0156] In some embodiments, during the process of performing signal change analysis on the first echo signal and the second echo signal to obtain the first characteristic parameter corresponding to the target cell, the processor 301 executes:

[0157] Determining a target window and a target step size, and cutting the first echo signal according to the target window and the target step size to obtain a plurality of first window signals;

[0158] Cutting the second echo signal according to the target window and the target step size to obtain a plurality of second window signals;

[0159] Acquire multiple target window groups by making a one-to-one correspondence between a plurality of the first window signals and a plurality of the second window signals, each of the target window groups including a first window signal and a second window signal corresponding to the first window signal;

[0160] Sequencing the first window signal and the second window signal in the target window group respectively to obtain first sequence information corresponding to the first window signal and second sequence information corresponding to the second window signal;

[0161] Determine the information change entropy corresponding to the target window group according to the first sequence information and the second sequence information;

[0162] A target curve is obtained by performing curve fitting according to the information change entropy, and the first characteristic parameter corresponding to the target cell is determined according to the target curve.

[0163] In some embodiments, during the process of determining the first characteristic parameter corresponding to the target cell according to the target curve, the processor 301 executes:

[0164] Obtaining a maximum value corresponding to the target curve, and determining a first moving direction corresponding to the target cell according to the maximum value;

[0165] Performing an integration operation on the target curve according to the target window and the target step length to obtain a target area corresponding to the target window;

[0166] Obtaining a first cell content corresponding to a unit entropy, and determining the number of cells corresponding to the target cells in the target window according to the target area and the first cell content;

[0167] Determine a first cell density corresponding to the target window according to the cell number;

[0168] The first characteristic parameter corresponding to the target cell is determined according to the first cell density and the first moving direction.

[0169] In some embodiments, during the process of performing image change analysis on the first image information and the second image information to obtain the second characteristic parameter corresponding to the target cell, the processor 301 executes:

[0170] Using the coding layer of the edge segmentation model to perform feature extraction on the first image information to obtain a corresponding first coding feature;

[0171] Using the edge perception layer of the edge segmentation model to extract edges of the first image information, to obtain corresponding first edge features;

[0172] Using the feature fusion layer of the edge segmentation model, the first coding feature and the first edge feature are superimposed to obtain a first enhanced feature;

[0173] Using the reordering layer of the edge segmentation model, feature fusion and channel rearrangement are performed according to the first encoding feature, the first edge feature and the first enhanced feature to obtain a first fused feature;

[0174] Determine a first target convolution kernel corresponding to the first fusion feature by using the adaptive channel attention layer of the edge segmentation model, and perform feature learning on the first fusion feature according to the first target convolution kernel to obtain a first target feature;

[0175] Determine first edge information corresponding to the first image information according to the first target feature using the edge recognition layer of the edge segmentation model;

[0176] Using the coding layer to extract features from the second image information to obtain corresponding second coding features;

[0177] Using the edge perception layer to perform edge extraction on the second image information to obtain corresponding second edge features;

[0178] Using the feature fusion layer to perform feature superposition on the second coding feature and the second edge feature to obtain a second enhanced feature;

[0179] Using the reordering layer to perform feature fusion and channel rearrangement according to the second coding feature, the second edge feature and the second enhanced feature, to obtain a second fused feature;

[0180] Determine a second target convolution kernel corresponding to the second fusion feature by using the adaptive channel attention layer, and perform feature learning on the second fusion feature according to the second target convolution kernel to obtain a second target feature;

[0181] Determining second edge information corresponding to the second image information according to the second target feature using the edge recognition layer;

[0182] An image change analysis is performed according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell.

[0183] In some embodiments, during the process of performing image change analysis according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell, the processor 301 executes:

[0184] Obtaining a first edge image corresponding to the first edge information and first position information corresponding to the first edge information from the first image information, and obtaining a second edge image corresponding to the second edge information and second position information corresponding to the second edge information from the second image information;

[0185] Performing pixel difference analysis on the first edge image to obtain a first salient area corresponding to the first edge image, and determining a first target position corresponding to the first salient area according to the first position information;

[0186] Performing pixel difference analysis on the second edge image to obtain a second salient area corresponding to the second edge image, and determining a second target position corresponding to the second salient area according to the second position information;

[0187] Determining a second moving direction corresponding to the target cell according to the first target position and the second target position;

[0188] determining a second cell content corresponding to a preset pixel, and obtaining a first pixel distribution corresponding to the second edge image;

[0189] Determining a second cell density corresponding to the target cells according to the second cell content and the first pixel distribution;

[0190] The second characteristic parameter corresponding to the target cell is determined according to the second cell density and the second moving direction.

[0191] In some implementations, the processor 301, in the process of performing pixel difference analysis on the first edge image to obtain the first salient area corresponding to the first edge image, executes:

[0192] Using a pixel segmentation layer of a salient information extraction model to perform region segmentation according to pixel differences between the first edge images to obtain a plurality of first segmented images;

[0193] Using the information extraction layer of the salient information extraction model to perform multi-channel feature extraction on the first segmented image to obtain a first feature map;

[0194] Using the information fusion layer of the salient information extraction model to perform feature fusion on the first feature map and the first position information to obtain a first fusion map;

[0195] Using the data processing layer of the salient information extraction model to perform data analysis on the first fusion image to obtain a first mean and a first standard deviation;

[0196] Using the salient feature extraction layer of the salient information extraction model to extract features according to the first mean and the first standard deviation in combination with the first fusion graph, to obtain a first salient feature;

[0197] Using the salient region determination layer of the salient information extraction model to perform saliency classification on the first segmented image according to the first salient feature, to obtain the first salient region corresponding to the first edge image;

[0198] The first significant feature is obtained according to the following formula:

[0199]

[0200] Among them, Xi represents the first significant feature corresponding to the i-th first segmented image, sigmoid represents the activation function, γ, θ and represents the training parameters corresponding to the salient feature extraction layer, α i represents the first mean corresponding to the i-th first segmented image, β i represents the first standard deviation corresponding to the i-th first segmented image, T i represents the i-th first fusion graph.

[0201] In some embodiments, during the process of fusing the first characteristic parameter and the second characteristic parameter to obtain the target characteristic parameter corresponding to the target cell, the processor 301 executes:

[0202] Determine a first weight corresponding to the first characteristic parameter and a second weight corresponding to the second characteristic parameter;

[0203] The first characteristic parameter and the second characteristic parameter are subjected to information fusion according to the first weight and the second weight to obtain the target characteristic parameter corresponding to the target cell.

[0204] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned embodiment of the cell invasion and metastasis ability monitoring method based on radio frequency imaging analysis, and will not be repeated here.

[0205] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis provided in the description of the embodiment of the present invention.

[0206] The storage medium may be an internal storage unit of the terminal device described in the foregoing embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc., equipped on the terminal device.

[0207] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0208] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0209] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis, characterized in that: The method comprises: In the induced transfer, the millimeter wave technology is used to continuously capture the echo signal of the lower chamber to obtain a first echo signal corresponding to the first moment and a second echo signal corresponding to the second moment of the lower chamber; Using an image sensor to collect first image information corresponding to the lower chamber at the first moment and second image information corresponding to the second moment, the lower chamber being a culture medium containing serum; Performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell; Performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell; fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell; The invasion and metastasis capabilities of the target cells are monitored according to the target characteristic parameters to obtain target detection results corresponding to the target cells.

2. The method according to claim 1, characterized in that The performing signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell includes: Determining a target window and a target step size, and cutting the first echo signal according to the target window and the target step size to obtain a plurality of first window signals; Cutting the second echo signal according to the target window and the target step size to obtain a plurality of second window signals; Acquire multiple target window groups by making a one-to-one correspondence between a plurality of the first window signals and a plurality of the second window signals, each of the target window groups including a first window signal and a second window signal corresponding to the first window signal; Sequencing the first window signal and the second window signal in the target window group respectively to obtain first sequence information corresponding to the first window signal and second sequence information corresponding to the second window signal; Determine the information change entropy corresponding to the target window group according to the first sequence information and the second sequence information; A target curve is obtained by performing curve fitting according to the information change entropy, and the first characteristic parameter corresponding to the target cell is determined according to the target curve.

3. The method according to claim 2, characterized in that The determining the first characteristic parameter corresponding to the target cell according to the target curve includes: Obtaining a maximum value corresponding to the target curve, and determining a first moving direction corresponding to the target cell according to the maximum value; Performing an integration operation on the target curve according to the target window and the target step length to obtain a target area corresponding to the target window; Obtaining a first cell content corresponding to a unit entropy, and determining the number of cells corresponding to the target cells in the target window according to the target area and the first cell content; Determine a first cell density corresponding to the target window according to the cell number; The first characteristic parameter corresponding to the target cell is determined according to the first cell density and the first moving direction.

4. The method according to claim 1, characterized in that The performing image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell includes: Using the coding layer of the edge segmentation model to perform feature extraction on the first image information to obtain a corresponding first coding feature; Using the edge perception layer of the edge segmentation model to extract edges of the first image information, to obtain corresponding first edge features; Using the feature fusion layer of the edge segmentation model, the first coding feature and the first edge feature are superimposed to obtain a first enhanced feature; Using the reordering layer of the edge segmentation model, feature fusion and channel rearrangement are performed according to the first encoding feature, the first edge feature and the first enhanced feature to obtain a first fused feature; Determine a first target convolution kernel corresponding to the first fused feature by using the adaptive channel attention layer of the edge segmentation model, and perform feature learning on the first fused feature according to the first target convolution kernel to obtain a first target feature; Determine first edge information corresponding to the first image information according to the first target feature using the edge recognition layer of the edge segmentation model; Using the coding layer to extract features from the second image information to obtain corresponding second coding features; Using the edge perception layer to perform edge extraction on the second image information to obtain corresponding second edge features; Using the feature fusion layer to perform feature superposition on the second coding feature and the second edge feature to obtain a second enhanced feature; Using the reordering layer to perform feature fusion and channel rearrangement according to the second coding feature, the second edge feature and the second enhanced feature, to obtain a second fused feature; Determine a second target convolution kernel corresponding to the second fusion feature by using the adaptive channel attention layer, and perform feature learning on the second fusion feature according to the second target convolution kernel to obtain a second target feature; Determining second edge information corresponding to the second image information according to the second target feature using the edge recognition layer; An image change analysis is performed according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell.

5. The method according to claim 4, characterized in that The performing image change analysis according to the first edge information and the second edge information to obtain the second characteristic parameter corresponding to the target cell includes: Obtaining a first edge image corresponding to the first edge information and first position information corresponding to the first edge information from the first image information, and obtaining a second edge image corresponding to the second edge information and second position information corresponding to the second edge information from the second image information; Performing pixel difference analysis on the first edge image to obtain a first salient area corresponding to the first edge image, and determining a first target position corresponding to the first salient area according to the first position information; Performing pixel difference analysis on the second edge image to obtain a second salient area corresponding to the second edge image, and determining a second target position corresponding to the second salient area according to the second position information; Determining a second moving direction corresponding to the target cell according to the first target position and the second target position; determining a second cell content corresponding to a preset pixel, and obtaining a first pixel distribution corresponding to the second edge image; Determining a second cell density corresponding to the target cells according to the second cell content and the first pixel distribution; The second characteristic parameter corresponding to the target cell is determined according to the second cell density and the second moving direction.

6. The method according to claim 5, characterized in that The performing pixel difference analysis on the first edge image to obtain a first salient area corresponding to the first edge image includes: Using a pixel segmentation layer of a salient information extraction model to perform region segmentation according to pixel differences between the first edge images to obtain a plurality of first segmented images; Using the information extraction layer of the salient information extraction model to perform multi-channel feature extraction on the first segmented image to obtain a first feature map; Using the information fusion layer of the salient information extraction model to perform feature fusion on the first feature map and the first position information to obtain a first fusion map; Using the data processing layer of the salient information extraction model to perform data analysis on the first fusion image to obtain a first mean and a first standard deviation; Using the salient feature extraction layer of the salient information extraction model to extract features according to the first mean and the first standard deviation in combination with the first fusion graph, to obtain a first salient feature; Using the salient region determination layer of the salient information extraction model to perform saliency classification on the first segmented image according to the first salient feature, to obtain the first salient region corresponding to the first edge image; The first significant feature is obtained according to the following formula: Among them, X i represents the first significant feature corresponding to the i-th first segmented image, sigmoid represents the activation function, γ, θ and represents the training parameters corresponding to the salient feature extraction layer, α i represents the first mean corresponding to the i-th first segmented image, β i represents the first standard deviation corresponding to the i-th first segmented image, T i represents the i-th first fusion graph.

7. The method according to claim 1, characterized in that The fusing the first characteristic parameter and the second characteristic parameter to obtain the target characteristic parameter corresponding to the target cell includes: Determine a first weight corresponding to the first characteristic parameter and a second weight corresponding to the second characteristic parameter; The first characteristic parameter and the second characteristic parameter are subjected to information fusion according to the first weight and the second weight to obtain the target characteristic parameter corresponding to the target cell.

8. A cell invasion and metastasis ability monitoring system based on radio frequency image analysis, characterized in that: include: A signal acquisition module, used to continuously capture the echo signal of the lower chamber using millimeter wave technology during induced transfer, and obtain a first echo signal corresponding to the lower chamber at a first moment and a second echo signal corresponding to the second moment; An image acquisition module, used for acquiring, by using an image sensor, first image information corresponding to the first moment and second image information corresponding to the second moment of the lower chamber, wherein the lower chamber is a culture medium containing serum; A first analysis module, configured to perform signal change analysis on the first echo signal and the second echo signal to obtain a first characteristic parameter corresponding to the target cell; A second analysis module, configured to perform image change analysis on the first image information and the second image information to obtain a second characteristic parameter corresponding to the target cell; A parameter fusion module, used for fusing the first characteristic parameter and the second characteristic parameter to obtain a target characteristic parameter corresponding to the target cell; The result determination module is used to monitor the invasion and metastasis ability of the target cells according to the target characteristic parameters to obtain the target detection results corresponding to the target cells.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program and implement the method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis according to any one of claims 1 to 7 when executing the computer program.

10. A computer storage medium for computer storage, characterized in that: The computer storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the method for monitoring cell invasion and metastasis ability based on radio frequency imaging analysis as described in any one of claims 1 to 7.

Images