A mixed hepatocyte terahertz quantitative detection system, electronic equipment and medium

Abstract

This invention discloses a mixed hepatocyte terahertz quantitative detection system, electronic device, and medium, comprising: a spectral preprocessing module for acquiring a mixed hepatocyte time-domain spectrum and converting it into a mixed hepatocyte terahertz frequency-domain spectrum using a fast Fourier transform; calculating the refractive index and absorption coefficient to obtain frequency-domain signals of the refractive index and absorption coefficient, and processing them using Gram angle sum field and Gram angle difference field to obtain Gram angle sum field map and Gram angle difference field map; an image feature extraction module for extracting gradient features from the Gram angle sum field map and Gram angle difference field map using gradient histograms; extracting grayscale features from the Gram angle sum field map and Gram angle difference field map using grayscale histograms; fusing the gradient features and grayscale features to obtain image features; and a hepatocyte cancer cell content quantitative detection module for inputting the image features into a pre-trained hepatocyte cancer cell content quantitative detection model to obtain the proportion of hepatocyte cancer cells.

Description

Technical Field

[0001] This invention relates to the field of spectral analysis and detection, specifically to a hybrid hepatocyte terahertz quantitative detection system, electronic device, and medium. Background Technology

[0002] Cancer threatens human life and health, and identifying malignant tumors through cell detection is a research direction of significant clinical importance. However, traditional spectral analysis methods have limitations, struggling to analyze nonlinear relationships and failing to consider data correlation. Furthermore, the environment for cell detection is a complex mixture, necessitating consideration of how to achieve quantitative cell detection in such environments.

[0003] Therefore, there is an urgent need to propose a mixed hepatocyte quantitative detection scheme to better achieve specific, non-destructive, and non-invasive mixed hepatocyte quantitative detection. Summary of the Invention

[0004] In view of this, the present invention provides a hybrid hepatocyte terahertz quantitative detection system, electronic device, and medium.

[0005] In a first aspect, embodiments of the present invention provide a hybrid hepatocyte terahertz quantitative detection system, the system comprising:

[0006] The spectral preprocessing module is used to acquire the time-domain spectrum of mixed hepatocytes and convert it into a terahertz frequency-domain spectrum of mixed hepatocytes through fast Fourier transform; it calculates the refractive index and absorption coefficient to obtain the frequency-domain signals of the refractive index and absorption coefficient; it processes the frequency-domain signals of the refractive index and absorption coefficient through Gram angle sum field and Gram angle difference field to obtain the Gram angle sum field map and Gram angle difference field map.

[0007] The image feature extraction module is used to extract gradient features from the Gram angle and field map and the Gram angle difference field map respectively through gradient histograms, and perform dimensionality reduction based on principal component analysis; it also extracts grayscale features from the Gram angle and field map and the Gram angle difference field map respectively through grayscale histograms, and performs dimensionality reduction based on principal component analysis; finally, it performs feature fusion on the gradient features and grayscale features to obtain image features.

[0008] The liver cancer cell content quantitative detection module is used to input image features into a pre-trained liver cancer cell content quantitative detection model to obtain the proportion of liver cancer cells.

[0009] Secondly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to realize the above-described hybrid hepatocyte terahertz quantitative detection system.

[0010] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described mixed hepatocyte terahertz quantitative detection system.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides a mixed hepatocyte terahertz quantitative detection system. The system preprocesses the mixed hepatocyte terahertz frequency domain spectrum using a Gram angle field through a spectral preprocessing module, extracts local gradient features using a gradient histogram, extracts global gray-level distribution features using a gray-level histogram, obtains image features, and performs quantitative analysis of hepatocyte content using a support vector machine regression model, thereby realizing the quantitative detection of hepatocytes in the cell mixture. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of a mixed hepatocyte terahertz quantitative detection system provided in an embodiment of the present invention;

[0014] Figure 2 The refractive index and absorption coefficient tangent characteristic parameter spectra of HepG2 and MIHA mixed cell suspensions with different cell ratios provided in the embodiments of the present invention;

[0015] Figure 3 GASF and GADF diagrams of the refractive index of cell suspension provided for embodiments of the present invention;

[0016] Figure 4 A schematic diagram of the gradient and gradient direction of the refractive index of the cell suspension corresponding to GASF, provided for embodiments of the present invention;

[0017] Figure 5 Gray-scale histograms of GASF and GADF corresponding to the absorption coefficients of two different ratios of mixed cell suspensions provided in this embodiment of the invention;

[0018] Figure 6 A schematic diagram illustrating the quantitative detection results of a mixed cell suspension test set of HepG2 and MIHA provided in an embodiment of the present invention;

[0019] Figure 7 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] It should be noted that, unless otherwise specified, the features in the following embodiments and implementation methods can be combined with each other.

[0022] In this embodiment of the invention, a liquid culture tank purchased from Hellma Analytic, Germany, was used for cell experiments. The liver cancer cell lines Huh-7 and HepG2, and the normal liver cell line MIHA were cultured in DMEM medium (supplemented with 10% fetal bovine serum, 50 μg / mL streptomycin, and 50 μg / mL penicillin) at 37°C and 5% CO2 concentration. After digestion with trypsin-EDTA digestion solution (0.25%), 10 cells were collected using a cell counter. 5 Three different cell suspensions were prepared at cell / mL ratios, and mixed according to different proportions of hepatocellular carcinoma cells and normal hepatocytes. Cells were mixed using two methods: Huh-7 and MIHA, and HepG2 and MIHA. The proportions of hepatocellular carcinoma cells in the mixed cell suspensions were 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9, with a total cell concentration of 10. 5 The cell / mL ratio was 19 different mixed cell suspensions. Twelve samples were taken from each cell suspension, and each sample was measured five times, resulting in a total of 1140 sets of time-domain spectral data.

[0023] like Figure 1 As shown in the figure, this embodiment of the invention provides a hybrid hepatocyte terahertz quantitative detection system, the system comprising: a spectral preprocessing module, an image feature extraction module, and a hepatocyte cancer cell content quantitative detection module.

[0024] The spectral preprocessing module is used to acquire the time-domain spectrum of mixed hepatocytes and convert it into a terahertz frequency-domain spectrum of mixed hepatocytes using fast Fourier transform (FFT). It then calculates the refractive index and absorption coefficient to obtain their frequency-domain signals. Finally, it processes the frequency-domain signals of the refractive index and absorption coefficient using Gramian angular summation field (GASF) and Gramian angular difference field (GADF) to obtain the Gramian angular summation field map and Gramian angular difference field map, respectively.

[0025] Furthermore, the expressions for calculating the refractive index n(ω) and the absorption coefficient α(ω) are as follows:

[0026]

[0027]

[0028] In the formula, d is the optical path length of the liquid cell; c is the speed of light in a vacuum; ω is the frequency; n1(ω) is the refractive index of the transparent window of the liquid cell; and k(ω) is the extinction coefficient. ρ(ω) and ρ(ω) are the phase difference and amplitude ratio between the sampled signal and the reference signal.

[0029] like Figure 2 As shown, Figure 2 (a) shows the refractive index of the mixed HepG2 and MIHA cell suspensions at liver cancer cell proportions of 0–0.9. Figure 2 (b) shows the uptake coefficients of the mixed HepG2 and MIHA cell suspensions when the proportion of liver cancer cells was 0-0.9.

[0030] Furthermore, in this example, the similarity or mutual interference of wavelengths of different substances can lead to spectral overlap and distortion. To account for nonlinear relationships and the correlation of data sequences, the frequency domain signals of refractive index and absorption coefficient are processed using Gram angle sum field and Gram angle difference field to obtain Gram angle sum field and Gram angle difference field maps, including:

[0031] The frequency domain signals of refractive index and absorption coefficient are transformed from the rectangular coordinate system to the polar coordinate system, as shown in the following expressions:

[0032] φ=arccos(x i -1≤x i ≤1,x i ∈X

[0033]

[0034] In the formula, the frequency domain signals of the refractive index and absorption coefficient are a one-dimensional sequence X = {x1, x2, ..., x...} n}, t i This represents a timestamp, where N is the total length of the timestamp.

[0035] The expressions for obtaining the Gram angle sum field based on the cosine function and the Gram angle difference field based on the sine function are as follows:

[0036]

[0037]

[0038]

[0039]

[0040] In the formula, φ j is the polar coordinate angle of the data point, j∈[1,n], where n is the number of data points. It is the normalization of a one-dimensional sequence X. It is a sequence The transpose of .

[0041] like Figure 3 As shown, Figure 3 (a) in the figure represents the Gram angle and field diagram of the refractive index of the MIHA mixed cell suspension. Figure 3 (b) in the figure is the Gram angular difference field plot of the refractive index of the MIHA mixed cell suspension. Figure 3 (c) in the figure represents the Gram angle and field diagram of the refractive index of a cell suspension in which HepG2 and MIHA are mixed at a ratio of 0.5. Figure 3 (d) in the figure is the Gram angular difference field plot of the refractive index of a cell suspension in which HepG2 and MIHA are mixed in a ratio of 0.5. Figure 3 (a) and Figure 3 In the diagram, (b) represents the Gram matrix when the mixing ratio is 0, i.e., only normal hepatocytes are present. Figure 3 (c) and Figure 3 In the figure, (d) represents the Gram matrix when the mixing ratio is 0.5, that is, when the concentrations of liver cancer cells and normal liver cells are the same.

[0042] The image feature extraction module is used to extract gradient features from the Gram angle and field map and the Gram angle difference field map using the Histogram of Oriented Gradient (HOG), and perform dimensionality reduction based on principal component analysis; it also extracts grayscale features from the Gram angle and field map and the Gram angle difference field map using the Grayscale Histogram (GH), and performs dimensionality reduction based on principal component analysis (PCA); finally, it fuses the gradient features and grayscale features to obtain the image features.

[0043] It should be noted that the present invention uses a combination of gradient histogram and grayscale histogram. Gradient histogram has more advantages in removing interference information from dopants and capturing signal differences in different cell proportions. Combined with the characteristics of the amplitude of each frequency point of the signal extracted by grayscale histogram, better detection results are obtained.

[0044] Specifically, the image feature extraction module uses histogram of gradients (HOG) to extract local gradient change features from the spectral correlation image of the terahertz signal. Figure 4 ), Figure 4 The gradient and gradient direction of images of two mixed cell suspensions with different ratios are shown. Figure 4 In the diagram, (a) represents the gradient and direction of the refractive index of the MIHA cell suspension corresponding to GASF. Figure 4 In Figure (b), the gradient and gradient direction of the refractive index of the cell suspension with a HepG2 and MIHA mixture ratio of 0.5 correspond to the GASF gradient. The HOG algorithm is used to accumulate the gradients in each direction and add them to the corresponding angle array to form a feature descriptor. Principal Component Analysis (PCA) is then used to reduce the dimensionality of the high-dimensional feature extraction results to 20 dimensions.

[0045] Furthermore, in this example, the Gram angle and field map and the Gram angle difference field map are divided into cell units of the same size and without overlap, and the gradient information of each pixel in each cell unit is collected; each cell unit is recorded as a block, and the gradient direction of each block is divided into several histogram channels on an average basis. The gradient of each pixel is accumulated and added to the corresponding angle array to form a feature descriptor vector; the Gram angle and field map and the Gram angle difference field map are traversed in the horizontal and vertical directions by the detection window in blocks to obtain gradient features.

[0046] For example, for a 100×100 grayscale image that has undergone spectral preprocessing, the gradient of each pixel is calculated, and then the gradient magnitude and gradient direction are calculated. The entire image is then divided into non-overlapping cell units of the same size, using 4×4 cell units. Gradient information of each pixel within these 4×4 cell units is collected, and these 4×4 cell units are grouped into a block. The gradient direction of each block is then divided into 9 histogram channels, and the gradients of each pixel are accumulated and added to the corresponding angle array to form a feature descriptor vector. The block is then scanned horizontally and vertically in the form of a detection window to traverse the image, obtaining the overall HOG features of the image.

[0047] Specifically, in this example, a grayscale histogram is used to collect global grayscale intensity distribution information, and the features are described by calculating the overall grayscale distribution. Figure 5 ), Figure 5 This is a gray-level histogram corresponding to the absorption coefficients of two cell suspensions mixed in different proportions. Figure 5 In the image, (a) is the grayscale histogram calculated by GASF. Figure 5 In the diagram, (b) shows the grayscale histogram calculated by GADF. Principal component analysis (PCA) was used to reduce the dimensionality of the high-dimensional feature extraction results to 20 dimensions.

[0048] The gray-level histogram is a function describing the global distribution of image gray levels. It is a statistical representation of the proportion of each gray level and can be expressed as:

[0049]

[0050]

[0051] In the formula, n k Let k be the number of pixels with a grayscale value of k, n be the total number of pixels in the image, and L be the number of grayscale levels.

[0052] After converting the grayscale image into a 256-level grayscale image, a grayscale histogram is calculated to supplement the overall brightness information of the image. The grayscale histogram is then used to directly statistically analyze the overall grayscale distribution information.

[0053] The liver cancer cell content quantitative detection module is used to input image features into a pre-trained liver cancer cell content quantitative detection model to obtain the proportion of liver cancer cells, realizing the quantitative detection of mixed liver cells with different proportions. Figure 6 ).

[0054] The quantitative detection model for liver cancer cell content adopts a support vector machine (SVR) regression model.

[0055] Furthermore, the training process of the quantitative detection model for liver cancer cell content includes:

[0056] All image features are divided into training and testing sets; in this example, 80% of the data for each concentration ratio is selected as the training set and 20% of the data is selected as the testing set.

[0057] The training set is input into the quantitative detection model of liver cancer cell content, which outputs the mixed proportion of liver cells to train the quantitative detection model of liver cancer cell content.

[0058] The test set was input into the liver cancer cell content quantification model, and the root mean square error (RMSE) and coefficient of determination R were used. 2 A quantitative detection model for assessing liver cancer cell content.

[0059] Accordingly, this application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the hybrid hepatocyte terahertz quantitative detection system as described above. Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities, including the hybrid hepatocyte terahertz quantitative detection system provided in this embodiment of the invention. (Except for...) Figure 7In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0060] Accordingly, this application also provides a computer-readable storage medium storing computer instructions thereon, which, when executed by a processor, implement the hybrid hepatocyte terahertz quantitative detection system as described above. The computer-readable storage medium can be an internal storage unit of any data processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0061] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0062] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A hybrid hepatocyte terahertz quantitative detection system, characterized in that, The system includes: The spectral preprocessing module is used to acquire the time-domain spectrum of mixed hepatocytes and convert it into a terahertz frequency-domain spectrum of mixed hepatocytes through fast Fourier transform; it calculates the refractive index and absorption coefficient to obtain the frequency-domain signals of the refractive index and absorption coefficient; it processes the frequency-domain signals of the refractive index and absorption coefficient through Gram angle sum field and Gram angle difference field to obtain the Gram angle sum field map and Gram angle difference field map. The image feature extraction module is used to extract gradient features from the Gram angle and field map and the Gram angle difference field map respectively through gradient histograms, and perform dimensionality reduction based on principal component analysis; it also extracts grayscale features from the Gram angle and field map and the Gram angle difference field map respectively through grayscale histograms, and performs dimensionality reduction based on principal component analysis; finally, it performs feature fusion on the gradient features and grayscale features to obtain image features. The liver cancer cell content quantitative detection module is used to input image features into a pre-trained liver cancer cell content quantitative detection model to obtain the proportion of liver cancer cells.

2. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The expressions for calculating the refractive index n(ω) and the absorption coefficient α(ω) are as follows: In the formula, d is the optical path length of the liquid cell; c is the speed of light in a vacuum; ω is the frequency; n1(ω) is the refractive index of the transparent window of the liquid cell; and k(ω) is the extinction coefficient. ρ(ω) and ρ(ω) are the phase difference and amplitude ratio between the sampled signal and the reference signal.

3. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, By processing the frequency domain signals of refractive index and absorption coefficient using Gram angle sum field and Gram angle difference field, the following Gram angle sum field and Gram angle difference field diagrams are obtained: The frequency domain signals of refractive index and absorption coefficient are transformed from the rectangular coordinate system to the polar coordinate system, as shown in the following expressions: φ=arccos(x i ),-1≤x i ≤1,x i ∈X In the formula, the frequency domain signals of the refractive index and absorption coefficient are a one-dimensional sequence X = {x1, x2, ..., x...} n }, t i This represents a timestamp, where N is the total length of the timestamp. The expressions for obtaining the Gram angle sum field based on the cosine function and the Gram angle difference field based on the sine function are as follows: In the formula, φ j is the polar coordinate angle of the data point, j∈[1,n], where n is the number of data points. It is the normalization of a one-dimensional sequence X. It is a sequence The transpose of .

4. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The gradient features of the Gram angle and field map, and the Gram angle difference field map were extracted using gradient histograms, including: The Gram angle and field map, and the Gram angle difference field map are divided into non-overlapping cell units of the same size. Gradient information of each pixel in each cell unit is collected. Each cell unit is recorded as a block. The gradient direction of each block is divided into several histogram channels. The gradient of each pixel is accumulated and added to the corresponding angle array to form a feature descriptor vector. The Gram angle and field map and the Gram angle difference field map are traversed in the horizontal and vertical directions by sliding the detection window in blocks to obtain gradient features.

5. The mixed hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The grayscale features of the Gram angle and field map, and the Gram angle difference field map were extracted from the grayscale histogram, including: The Gram angle and field map, and the Gram angle difference field map are converted into 256-level grayscale images. Grayscale distribution information is statistically analyzed through grayscale histograms, and grayscale features are obtained based on the grayscale distribution information.

6. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The expression for calculating gray-level distribution information using a gray-level histogram is as follows: In the formula, n k Let k be the number of pixels with a grayscale value of k, n be the total number of pixels in the image, and L be the number of grayscale levels.

7. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The quantitative detection model for liver cancer cell content adopts a support vector machine regression model.

8. The hybrid hepatocyte terahertz quantitative detection system according to claim 1, characterized in that, The training process of the quantitative detection model for liver cancer cell content includes: All image features are divided into training and testing sets; The training set is input into the quantitative detection model of liver cancer cell content, which outputs the mixed proportion of liver cells to train the quantitative detection model of liver cancer cell content. The test set was input into the quantitative detection model for liver cancer cell content, and the root mean square error and coefficient of determination were used to evaluate the quantitative detection model for liver cancer cell content.

9. An electronic device comprising a memory and a processor, characterized in that, The memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the hybrid hepatocyte terahertz quantitative detection system according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the mixed hepatocyte terahertz quantitative detection system as described in any one of claims 1-8.

Images