A method and system for ultrasound examination of thyroid nodules

Abstract

This invention relates to the field of ultrasound imaging analysis technology, specifically to a method and system for ultrasound examination of thyroid nodules, comprising the following steps: acquiring ultrasound echo signals of thyroid tissue, analyzing the instantaneous characteristics of the nodule region signal, calculating energy distribution and instantaneous frequency difference, screening stable resonance modes, and establishing a resonance mode feature matrix. In this invention, by optimizing the instantaneous feature analysis of ultrasound signals, the decomposition accuracy of echo data and the accurate identification capability of target signals are improved. Combined with energy distribution and frequency difference screening, background noise interference is significantly reduced. Image enhancement, through gradient change rate and transform kernel matrix optimization, improves the clarity of image details. Short-time frequency feature extraction and frequency shift measurement enhance the ability to identify lesion tissue. Simultaneously, through the analysis of frequency shift stability parameters and shift change rate, the accuracy of automatic nodule nature determination is improved, making ultrasound examination of thyroid nodules more scientific and accurate.

Description

Technical Field

[0001] This invention relates to the field of ultrasound imaging analysis technology, and in particular to a method and system for ultrasound examination of thyroid nodules. Background Technology

[0002] The field of ultrasound imaging analysis technology encompasses ultrasound imaging and data processing methods in medical imaging. Its core content involves the acquisition, transmission, imaging, and analysis of ultrasound signals, including the design and application of ultrasound transducers, the propagation characteristics of ultrasound signals, the reception and processing of echo signals, and image reconstruction and analysis methods. In medical diagnosis, ultrasound imaging is widely used for the visual detection of organs and tissues, enabling real-time, non-invasive examination through ultrasound signals of different frequencies and modes. Systematic research in this field covers the construction of ultrasound imaging equipment, ultrasound transmission and reception modes, tissue echo signal characteristic analysis, image enhancement and denoising, automatic target area identification and classification, and, in particular, the optimization of imaging quality using specific parameters based on the acoustic characteristics of different tissue structures, combined with computer-aided technology for automated data analysis and identification.

[0003] Thyroid nodule ultrasound examination refers to the use of ultrasound imaging technology to identify and classify the structure, morphology, boundaries, internal echo characteristics, and blood flow status of thyroid nodules. High-frequency ultrasound signals are used to acquire echo data of thyroid tissue, and the amplitude, phase, and spectral characteristics of the echo signals are extracted and analyzed to identify the echo type and tissue composition of the nodules. This includes analysis of nodule boundary features to determine its outline integrity, morphological regularity, and relationship with surrounding tissues. Simultaneously, blood flow signal analysis technology is combined with Doppler ultrasound to detect the distribution of blood vessels inside and outside the nodule. Finally, image feature matching and classification criteria are used to determine the imaging characteristics of thyroid nodules, providing a basis for subsequent clinical evaluation.

[0004] Current signal analysis technologies primarily rely on the amplitude and phase information of traditional echo signals, failing to adequately consider the impact of instantaneous frequency changes. This results in low signal decomposition accuracy and susceptibility to interference from tissue scattering effects. During signal enhancement, background noise suppression strategies are relatively simplistic and lack dynamic adjustment tailored to different tissue characteristics, potentially leading to insufficient signal enhancement or loss of the target signal. Image optimization relies on conventional contrast enhancement and edge sharpening techniques, making it difficult to effectively distinguish low-contrast nodule boundaries and impacting nodule morphological analysis. Frequency shift measurements fail to fully incorporate spatial location information, potentially missing local abnormal frequency shift signals and reducing the accuracy of lesion identification. The lack of analysis on the spatial distribution and stability parameters of abnormal regions makes nodule characterization dependent on subjective experience, increasing diagnostic uncertainty. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a method and system for ultrasound examination of thyroid nodules.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for ultrasound examination of thyroid nodules, comprising the following steps:

[0007] S1: Obtain ultrasound echo signals from thyroid tissue, analyze the instantaneous characteristics of the nodule region signal, calculate the energy distribution and instantaneous frequency difference, screen stable resonance modes, and establish a resonance mode feature matrix;

[0008] S2: Call the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculate the gain adjustment coefficient based on the signal peak curve, perform nonlinear mapping based on the gain adjustment coefficient, call the cross-correlation matrix to filter the correlated signals, perform dynamic weight suppression to adjust the signal strength, and obtain the enhanced echo signal matrix;

[0009] S3: Call the enhanced echo signal matrix, calculate the gray-level gradient of the nodule region, adjust the pixel contrast, and obtain the super-resolution enhanced image matrix;

[0010] S4: Call the super-resolution enhanced image matrix, extract the short-time frequency features of the nodule region, calculate the frequency change, and obtain the echo frequency offset matrix;

[0011] S5: Call the echo frequency offset matrix to calculate the abnormal coverage ratio and frequency shift stability, and obtain the ultrasound imaging analysis steps.

[0012] As a further aspect of the present invention, the resonant mode feature matrix includes modal energy characteristics, resonant frequency distribution, and instantaneous amplitude parameters; the enhanced echo signal matrix includes signal gain adjustment coefficient, noise suppression weight, and target signal energy ratio; the super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness equalization adjustment coefficients; the echo frequency offset matrix includes spatial frequency shift amplitude, local frequency anomaly index, and main frequency peak change rate; and the ultrasound imaging analysis steps include abnormal area coverage, frequency offset stability, and nodule tissue frequency shift characteristics.

[0013] As a further aspect of the present invention, the specific steps for obtaining ultrasound echo signals of thyroid tissue, analyzing the instantaneous characteristics of the nodule region signal, calculating the energy distribution and instantaneous frequency difference, screening stable resonance modes, and establishing a resonance mode feature matrix are as follows:

[0014] S101: Acquire ultrasound echo signals of thyroid tissue, analyze the instantaneous amplitude, phase and short-time frequency of the signal, calculate the amplitude change rate, phase shift and short-time frequency change within the time window, and obtain the instantaneous signal feature matrix;

[0015] S102: Based on the instantaneous signal feature matrix, calculate the energy distribution of the frequency band signal, obtain the signal energy of the differentiated frequency bands, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency bands, calculate the frequency change rate, analyze the frequency difference distribution between frequency bands, determine the trend of frequency difference in the time dimension, and obtain the instantaneous frequency difference matrix.

[0016] S103: Call the instantaneous frequency difference matrix, filter the resonant modes with stable frequency differences, extract the signal modes that meet the threshold requirements, integrate the corresponding amplitude, phase and short-time frequency data, and establish the resonant mode feature matrix.

[0017] As a further aspect of the present invention, the specific steps for calling the resonant mode feature matrix, calculating the modal energy density of the echo signal in the nodule region, calculating the gain adjustment coefficient based on the signal peak curve, performing nonlinear mapping based on the gain adjustment coefficient, calling the cross-correlation matrix to filter correlated signals, and performing dynamic weight suppression to adjust the signal strength to obtain the enhanced echo signal matrix are as follows:

[0018] S201: Call the resonant mode feature matrix to analyze the modal energy density of the echo signal in the nodule region, extract the energy value of the frequency band signal, calculate the average energy value per unit area, and statistically analyze the energy fluctuation range to obtain the modal energy peak value and corresponding frequency, and obtain the modal energy density matrix.

[0019] S202: Based on the modal energy density matrix, analyze the gain adjustment coefficient of the signal peak curve, call the energy change rate analysis of the peak region to analyze the gain adjustment factor, establish a nonlinear mapping function based on the energy gradient, adjust the signal intensity, calculate the cross-correlation value of the background organization modal signals, construct the cross-correlation matrix between signals, filter signals that meet the correlation threshold, adjust the signal mapping parameters based on the correlation coefficient, filter the background signals that meet the constraints, and obtain the signal cross-correlation matrix.

[0020] S203: Call the aforementioned signal cross-correlation matrix, calculate the signal cross-correlation coefficient, calculate the dynamic suppression factor based on the correlation coefficient, adjust the signal strength, perform dynamic weighted suppression, integrate the adjusted signal matrix, and obtain the enhanced echo signal matrix.

[0021] As a further aspect of the present invention, the formula for calculating the average energy is specifically as follows:

[0022]

[0023] Among them, E avg The modal energy mean of a unit region is represented by N, where N represents the number of sampling points per unit region, and A represents the mean modal energy. i B represents the real part of the signal component at the i-th sampling point. iC represents the imaginary part of the signal at the i-th sampling point. i Represents the instantaneous frequency of the i-th sampling point. It represents the average instantaneous frequency of all sampling points within a unit area.

[0024] As a further aspect of the present invention, the specific steps for calling the enhanced echo signal matrix, calculating the gray-level gradient of the nodule region, adjusting pixel contrast, and obtaining the super-resolution enhanced image matrix are as follows:

[0025] S301: Call the enhanced echo signal matrix to calculate the gray-level gradient of the ultrasound image of the nodule region, extract the gray-level change rate of each pixel in the image, calculate the gradient change amplitude between pixels based on the gradient distribution, construct the gray-level gradient distribution matrix, and obtain the gradient distribution matrix.

[0026] S302: Based on the gradient distribution matrix, analyze the gradient change rate, statistically analyze the gradient change trend of pixels in the image area, determine the transformation kernel parameters according to the gradient change rate, call the transformation kernel parameters to adjust the pixel contrast, calculate the contrast change value of the pixel, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix.

[0027] S303: Call the transformation kernel matrix, calculate the neighborhood brightness adjustment value in combination with the brightness gradient, extract the neighborhood brightness change trend of the pixel, identify the brightness change coefficient, correct the image pixel brightness according to the adjustment coefficient, integrate the adjusted image data, and obtain the super-resolution enhanced image matrix.

[0028] As a further aspect of the present invention, the gradient change trend of pixels within the statistical image area is determined by the following formula:

[0029]

[0030] The transformation kernel parameters are determined based on the gradient change rate. The transformation kernel parameters are then used to adjust the pixel contrast. The contrast change value of each pixel is calculated. The image contrast distribution is adjusted based on the change value to obtain the transformation kernel matrix.

[0031] Where ΔK represents the percentage of the average gradient change rate, M represents the total number of pixels in the computational region, and G... k G represents the gradient value of the k-th pixel. avg The summation sign represents the average gradient values ​​of all pixels within the region. This indicates that the gradient changes of all pixels are accumulated.

[0032] As a further aspect of the present invention, the specific steps for calling the super-resolution enhanced image matrix, extracting the short-time frequency features of the nodule region, calculating the frequency change, and obtaining the echo frequency offset matrix are as follows:

[0033] S401: Call the super-resolution enhanced image matrix, extract the short-time frequency features of the echo signal in the nodule region, calculate the instantaneous frequency value of the pixel, obtain the frequency change trend of the local region, identify the instantaneous frequency peak, and generate an instantaneous frequency peak matrix.

[0034] S402: Based on the instantaneous frequency peak matrix, calculate the frequency change of the pixel in the spatial position, extract the frequency change difference between pixels, count the rate of change in the local frequency gradient direction, calculate the spatial frequency offset amplitude value, analyze the local change trend of the offset distribution, construct the offset distribution matrix according to the frequency change rate, filter the regions with offset values ​​higher than a set threshold, and obtain the offset distribution matrix.

[0035] S403: Call the offset distribution matrix to analyze the change characteristics of the spatial frequency gradient, calculate the degree of local frequency drift, filter the frequency offset of abnormal areas, integrate the frequency change data, and obtain the echo frequency offset matrix.

[0036] As a further aspect of the present invention, the specific steps for obtaining the ultrasound imaging analysis steps are as follows: The echo frequency offset matrix is ​​invoked to calculate the abnormal coverage ratio and frequency shift stability.

[0037] S501: Call the echo frequency offset matrix, count the spatial distribution density of the abnormal frequency shift region inside the nodule, calculate the proportion of the number of abnormal frequency shift points in the region, calculate the spatial coverage of the abnormal region based on the local distribution, and obtain the abnormal frequency shift spatial density matrix.

[0038] S502: Based on the abnormal frequency shift spatial density matrix, calculate the coverage ratio of the abnormal region within the nodule range, statistically analyze the frequency shift value variation range of the covered region, calculate the frequency fluctuation rate of the abnormal region, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient based on the frequency shift mean of the differentiated region, filter regions with stability below the threshold, extract abnormal frequency shift stability features, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix.

[0039] S503: Call the tissue frequency shift stability matrix, calculate the spatial change of frequency shift trend, extract the frequency shift rate of differential regions, statistically analyze the distribution characteristics of the shift change regions, integrate regional frequency shift change data, and obtain ultrasound imaging analysis steps.

[0040] A thyroid nodule ultrasound examination system, comprising:

[0041] The resonant mode extraction module acquires ultrasound echo signals from thyroid tissue, analyzes the instantaneous amplitude, phase, and short-time frequency characteristics of the echo signals in the nodule region, calculates the energy distribution and instantaneous frequency difference, and establishes a resonant mode feature matrix.

[0042] The signal enhancement and control module calls the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping based on the gain adjustment coefficient, statistically analyzes the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor based on the signal cross-correlation coefficient, performs dynamic weight suppression to adjust the signal strength, and obtains the enhanced echo signal matrix.

[0043] The image gradient optimization module calls the enhanced echo signal matrix to calculate the gray-level gradient of the ultrasound image in the nodule region, constructs the gradient distribution matrix, calls the gradient change rate to calculate the transform kernel matrix, adjusts the pixel contrast according to the transform kernel matrix, calculates the neighborhood brightness adjustment value, and obtains the super-resolution enhanced image matrix.

[0044] The frequency offset detection module calls the super-resolution enhanced image matrix to extract the short-time frequency characteristics of the echo signal in the nodule region, calculates the instantaneous frequency peak, constructs an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to filter abnormal regions, and obtains the echo frequency offset matrix.

[0045] The spatial trend analysis module calls the echo frequency offset matrix to statistically analyze the spatial distribution density of abnormal frequency shift regions within the nodule, calculates the abnormal coverage ratio, calls the frequency shift trend matrix to calculate the offset change rate, and obtains the ultrasound imaging analysis steps.

[0046] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0047] In this invention, by optimizing the instantaneous feature analysis of ultrasound signals, the decomposition accuracy of echo data and the accurate identification capability of target signals are improved. Combined with energy distribution and frequency difference screening, background noise interference is significantly reduced. Image enhancement improves the clarity of image details through gradient change rate and transformation kernel matrix optimization. Short-time frequency feature extraction and frequency shift measurement enhance the ability to identify lesion tissue. At the same time, by analyzing frequency shift stability parameters and shift change rate, the accuracy of automatic determination of nodule nature is improved, making thyroid nodule ultrasound examination more scientific and accurate. Attached Figure Description

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

[0049] Figure 1 This is a schematic diagram of the steps of the present invention;

[0050] Figure 2This is a flowchart of steps S1 of the present invention;

[0051] Figure 3 This is a flowchart of steps S2 of the present invention;

[0052] Figure 4 This is a flowchart of steps S3 of the present invention;

[0053] Figure 5 This is a flowchart of step S4 of the present invention;

[0054] Figure 6 This is a flowchart of steps S5 of the present invention;

[0055] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0056] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0058] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0059] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0060] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0061] Please see Figure 1 A method for ultrasound examination of thyroid nodules includes the following steps:

[0062] S1: Acquire ultrasound echo signals of thyroid tissue, analyze the instantaneous amplitude, phase and short-time frequency characteristics of the echo signals in the nodule region, calculate the energy distribution and instantaneous frequency difference, use the frequency difference to screen stable resonance modes, and establish a resonance mode feature matrix;

[0063] S2: Call the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculate the gain adjustment coefficient based on the signal peak curve, perform nonlinear mapping based on the gain adjustment coefficient, statistically analyze the cross-correlation value of the background tissue modal signal, call the cross-correlation matrix to filter the correlated signal, calculate the dynamic suppression factor based on the signal cross-correlation coefficient, perform dynamic weight suppression to adjust the signal strength, and obtain the enhanced echo signal matrix.

[0064] S3: Call the enhanced echo signal matrix, calculate the gray-level gradient of the ultrasound image in the nodule region, construct the gradient distribution matrix, call the gradient change rate to calculate the transform kernel matrix, adjust the pixel contrast according to the transform kernel matrix, combine the brightness gradient to calculate the neighborhood brightness adjustment value, and obtain the super-resolution enhanced image matrix.

[0065] S4: Call the super-resolution enhanced image matrix to extract the short-time frequency characteristics of the echo signal in the nodule region, calculate the instantaneous frequency peak, call the peak matrix to calculate the frequency change in spatial location, construct the offset distribution matrix based on the frequency offset amplitude, call the offset distribution matrix to filter abnormal areas, and obtain the echo frequency offset matrix.

[0066] S5: Call the echo frequency offset matrix to statistically analyze the spatial distribution density of abnormal frequency shift regions within the nodule, calculate the abnormal coverage ratio, call the coverage ratio matrix to calculate tissue frequency shift stability, call the frequency shift trend matrix to calculate the offset change rate, and obtain the ultrasound imaging analysis steps.

[0067] The resonant mode feature matrix includes modal energy characteristics, resonant frequency distribution, and instantaneous amplitude parameters; the enhanced echo signal matrix includes signal gain adjustment coefficient, noise suppression weight, and target signal energy ratio; the super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness equalization adjustment coefficients; the echo frequency offset matrix includes spatial frequency shift amplitude, local frequency anomaly index, and dominant frequency peak change rate; and the ultrasound imaging analysis steps include abnormal area coverage, frequency offset stability, and nodule tissue frequency shift characteristics.

[0068] Please see Figure 2 The specific steps of S1 are as follows:

[0069] S101: Acquire ultrasound echo signals of thyroid tissue, analyze the instantaneous amplitude, phase and short-time frequency of the signal, calculate the amplitude change rate, phase shift and short-time frequency change within the time window, and obtain the instantaneous signal feature matrix;

[0070] Acquire ultrasound echo signals from thyroid tissue, collect data from the echo signals, and set the sampling frequency f. s The frequency is 40MHz, the sampling time T is 50ms, and the number of signal data points acquired is N = f s ×T=2×10 6 The signal data was bandpass filtered with upper and lower frequency limits set to 2MHz and 15MHz, respectively, retaining only the effective frequency band signal of the thyroid tissue echo. A Hilbert transform was then performed on the filtered signal to extract the instantaneous amplitude, instantaneous phase, and short-time frequency. The instantaneous amplitude A(t) was obtained by calculating the signal envelope, and the instantaneous phase φ(t) was obtained by taking the arctangent of the imaginary part of the Hilbert transform and the real part of the original signal. The short-time frequency f... i (t) is obtained by differentiating the phase with respect to time. A sliding window method is used to segment the signal, setting the window length to 512 points and the window sliding step size to 128 points. The amplitude change rate is calculated within each window, defined as dA / dt, and is obtained by dividing the amplitude difference between adjacent sampling points by the time step size Δt = 1 / f. s The phase offset is obtained and calculated, defined as the cumulative phase change Δφ = φ(t + Δt) - φ(t). The short-time frequency variation is calculated and defined as df / dt, obtained by dividing the short-time frequency difference between adjacent moments by the time step. These parameters are arranged in time order to form the instantaneous signal feature matrix.

[0071] S102: Based on the instantaneous signal feature matrix, calculate the energy distribution of the frequency band signal, obtain the signal energy of the differentiated frequency bands, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency bands, calculate the frequency change rate, analyze the frequency difference distribution between frequency bands, determine the changing trend of the frequency difference in the time dimension, and obtain the instantaneous frequency difference matrix.

[0072] Based on the instantaneous signal feature matrix, the energy distribution of signals in different frequency bands is calculated. A Short-Time Fourier Transform (STFT) is performed on the signal, using a Hanning window with a length of 512 points and a step size of 128 points. The calculated time-spectrum energy distribution matrix E(f,t) is then used to select frequency bands with higher energy, and an energy threshold E is set. th Add twice the standard deviation to the mean of the energy distribution, i.e., E th =μ E +2σ E Select the frequency bands that meet the threshold as the differentiated frequency bands, and calculate the instantaneous frequency f of the differentiated frequency bands. d (t), perform a difference operation on the instantaneous frequency to calculate the instantaneous frequency difference Δf = f between adjacent time points. d (t+Δt)-f d (t), calculate the rate of frequency change, defined as df d / dt is obtained by dividing the instantaneous frequency difference between adjacent moments by the time step, and the frequency difference distribution between different frequency bands is analyzed, with a time window T set. w =10ms, calculate the mean and standard deviation of the frequency difference within this time window, determine the trend of the frequency difference over time, use linear regression to fit the trend of the frequency difference, and statistically analyze the fitting slope k. If |k| is less than a set threshold k th If the frequency difference is 0.05 Hz / ms, then the frequency difference change in this frequency band is determined to be stable, and an instantaneous frequency difference matrix is ​​eventually formed.

[0073] S103: Call the instantaneous frequency difference matrix, filter the resonant modes with stable frequency differences, extract the signal modes that meet the threshold requirements, integrate the corresponding amplitude, phase and short-time frequency data, and establish the resonant mode feature matrix.

[0074] Resonant modes with stable frequency differences were selected, and statistical analysis was performed on each frequency band in the matrix to calculate the standard deviation σ of the frequency difference for each band over the entire signal duration. f Set threshold σ th σ is the median of the frequency differences. f <σ th The frequency band is used as the stable resonant mode. Signal modes that meet the threshold requirements are extracted. The amplitude, phase, and short-time frequency data of the corresponding frequency band are retrieved from the instantaneous signal feature matrix, and the amplitude threshold A is set. th A is 1.5 times the standard deviation of the mean amplitude. th =μ A +1.5σ A Screening amplitude greater than A th The signal points that meet the screening criteria are integrated to establish a resonance mode feature matrix, which ultimately forms a dataset for subsequent analysis.

[0075] Please see Figure 3 The specific steps of S2 are as follows:

[0076] S201: Call the resonant mode feature matrix to analyze the modal energy density of the echo signal in the nodule region, extract the energy value of the frequency band signal, calculate the average energy value per unit area, and statistically analyze the energy fluctuation range to obtain the modal energy peak value and corresponding frequency, and obtain the modal energy density matrix.

[0077] The formula for calculating the average energy is as follows:

[0078]

[0079] Among them, E avg The modal energy mean of a unit region is represented by N, where N represents the number of sampling points per unit region, and A represents the mean modal energy. iB represents the real part of the signal component at the i-th sampling point. i C represents the imaginary part of the signal at the i-th sampling point. i Represents the instantaneous frequency of the i-th sampling point. It represents the average instantaneous frequency of all sampling points within a unit area.

[0080] The formula is used to calculate the mean modal energy E per unit area. avg The amplitude and frequency information of the signal are used. For a series of sampling points i, the following steps are performed:

[0081] Calculate the complex signal amplitude at each sampling point Where A i and B i These are the real and imaginary components of the i-th sampling point, respectively. This step is calculated based on the complex representation of the signal, a common method in electrical signal processing for obtaining the instantaneous amplitude of the signal.

[0082] Calculate the frequency deviation at each sampling point Where C i It is the instantaneous frequency of the i-th sampling point, and This is the average instantaneous frequency of all sampling points within the entire sampling area. This step provides the degree of frequency variation at each point; a larger frequency deviation may indicate areas of concentrated energy or significant variation.

[0083] Summing and multiplying the two results above, and then dividing by the total number of sampling points N, yields the average energy E for the entire region. avg .

[0084] To perform this calculation, consider the following specific example:

[0085] Assume there are N=4 sampling points in a unit region, and the real and imaginary components and instantaneous frequency of each point are as follows:

[0086] A1 = 3, B1 = 4, C1 = 5Hz

[0087] A2 = 1, B2 = 1, C2 = 5.5 Hz

[0088] A3 = 2, B3 = 2, C3 = 4.5 Hz

[0089] A4 = 1, B4 = 3, C4 = 5Hz

[0090] First, calculate the average value of the instantaneous frequency:

[0091]

[0092] Calculate the amplitude and frequency deviation at each point and substitute them into the formula:

[0093]

[0094]

[0095] Finally, calculate E avg :

[0096]

[0097] The results show that the average energy density is 0.53 within the selected sampling area, which reflects the combined effect of the average energy state and frequency deviation of the signal within the area.

[0098] S202: Based on the modal energy density matrix, analyze the gain adjustment coefficient of the signal peak curve, call the energy change rate analysis of the peak region to analyze the gain adjustment factor, establish a nonlinear mapping function based on the energy gradient, adjust the signal intensity, calculate the cross-correlation value of the background organization modal signals, construct the cross-correlation matrix between signals, screen signals that meet the correlation threshold, adjust the signal mapping parameters based on the correlation coefficient, screen background signals that meet the constraints, and obtain the signal cross-correlation matrix.

[0099] Based on the modal energy density matrix, the peak energy density curve is analyzed, a gain adjustment coefficient is set, the energy change rate corresponding to the peak region is selected, the energy increment per unit time is calculated, and the gain adjustment factor is defined as the ratio of energy increment to standard energy. A nonlinear mapping function is calculated using the energy gradient, and the mapping parameter range is set, defining the energy gain adjustment range as between 0.8 and 1.2. The energy change gradient is discretely segmented, with each segment length set to 5% of the maximum energy change of the original signal. The corresponding gain parameters are calculated based on the gradient of different segments to adjust the signal strength. Cross-correlation calculations are performed on the modal signals of the background organization, a reference background signal region is defined, and the background signal is calculated. The cross-correlation values ​​of the signals and nodule signals are used to obtain the cross-correlation matrix between the signals. A threshold for the correlation coefficient is set, and 1.1 times the mean of the cross-correlation matrix is ​​used as the screening criterion to screen signals that meet the threshold range. The signal mapping parameters are adjusted according to the correlation coefficient. A correlation stratification standard is set, and the signals are divided into three groups according to the correlation coefficient from high to low. The range of the mapping parameters is adjusted so that the gain adjustment factor of the high correlation coefficient signal is between 0.9 and 1.1, the adjustment factor of the medium correlation coefficient signal is between 0.85 and 1.15, and the adjustment factor of the low correlation coefficient signal is between 0.8 and 1.2. The adjusted signals are normalized to form the signal cross-correlation matrix.

[0100] S203: Call the signal cross-correlation matrix, calculate the signal cross-correlation coefficient, calculate the dynamic suppression factor based on the correlation coefficient, adjust the signal strength, perform dynamic weighted suppression, integrate the adjusted signal matrix, and obtain the enhanced echo signal matrix;

[0101] The signal cross-correlation matrix is ​​invoked, and the signal data within the matrix are calculated. The cross-correlation coefficient between each pair of signals is calculated, and the mean of all cross-correlation coefficients is taken as a reference value. A dynamic suppression factor is set, and the cross-correlation coefficient of each signal is compared with the mean. If it is higher than the mean, its relative deviation is calculated, and the suppression factor is set to 0.7 times the deviation. If it is lower than the mean, the suppression factor is set to 1.2 times the deviation. The signal strength is adjusted by multiplying each signal data point by its corresponding suppression factor and performing dynamic weight suppression. The weight adjustment range is set so that the amplitude of the suppressed signal is kept between 70% and 120% of the original signal amplitude to avoid excessive attenuation or enhancement of the signal. All adjusted signal data are calculated and integrated to form an enhanced echo signal matrix.

[0102] Please see Figure 4 The specific steps of S3 are as follows:

[0103] S301: Call the enhanced echo signal matrix, calculate the gray-level gradient of the ultrasound image of the nodule region, extract the gray-level change rate of each pixel in the image, calculate the gradient change amplitude between pixels based on the gradient distribution, construct the gray-level gradient distribution matrix, and obtain the gradient distribution matrix.

[0104] The enhanced echo signal matrix is ​​invoked, and the image resolution is set to 512×512 pixels. Pixel-by-pixel calculations are performed on the image data to extract the grayscale value of each pixel. The grayscale gradient is defined as the difference between the grayscale values ​​of adjacent pixels. The Sobel operator is used for edge detection, calculating the gradient components in the horizontal and vertical directions. The window size is set to 3×3 pixels. The grayscale change rate between the center pixel and surrounding pixels is calculated. The gradient change rate is defined as the gradient change amplitude of each pixel divided by the time step, with the time step set to the image frame interval of 0.02 seconds. The distribution of the gradient change rate throughout the entire image area is statistically analyzed. The gradient values ​​are normalized, mapping the gradient amplitude to the range of 0 to 1. The gradient changes between different pixels are compared, and a grayscale gradient distribution threshold is set. 1.3 times the global grayscale gradient mean is taken as the standard, and areas exceeding this threshold are filtered out. The gradient data of all pixels are arranged to form a grayscale gradient distribution matrix, and finally, the gradient distribution matrix is ​​obtained.

[0105] S302: Based on the gradient distribution matrix, analyze the gradient change rate, statistically analyze the gradient change trend of pixels in the image area, determine the transform kernel parameters according to the gradient change rate, call the transform kernel parameters to adjust the pixel contrast, calculate the contrast change value of the pixel, adjust the image contrast distribution according to the change value, and obtain the transform kernel matrix.

[0106] The gradient change trend of pixels within the image region is statistically analyzed using the following formula:

[0107]

[0108] The transformation kernel parameters are determined based on the gradient change rate. The transformation kernel parameters are then used to adjust the pixel contrast. The contrast change value of each pixel is calculated. The image contrast distribution is adjusted based on the change value to obtain the transformation kernel matrix.

[0109] Where ΔK represents the percentage of the average gradient change rate, M represents the total number of pixels in the computational region, and G... k G represents the gradient value of the k-th pixel. avg The summation sign represents the average gradient values ​​of all pixels within the region. This indicates that the gradient changes of all pixels are accumulated;

[0110] The provided formula calculates the gradient change rate for each pixel in the image, expressed as a percentage change relative to the average gradient value. The formula first calculates the average gradient G of all pixels within the region. avg :

[0111]

[0112] Then, the gradient G for each pixel k Calculate its average gradient G avg The deviation, and the deviation relative to G avg Convert to percentages. This method efficiently depicts the intensity and directionality of pixel gradient changes.

[0113] Let's set a specific example: Assume there are M = 5 pixels in a computational region, and the gradient values ​​of each pixel are as follows: -G1 = 10, G2 = 12, G3 = 15, G4 = 11, G5 = 13

[0114] Calculate G avg :

[0115]

[0116] Substitute the values ​​into the formula to calculate the gradient rate of change for each pixel:

[0117]

[0118] The result shows that the average gradient change rate is 11.72% within a given region, reflecting the average intensity of gradient changes within that region. Parameter descriptions: ΔK represents the average gradient change rate, used to measure the average intensity of pixel gradient changes; M is the total number of pixels considered, used to determine the breadth of the calculation; G... k It is the gradient value of a single pixel, obtained directly from the image processing algorithm, G avgIt is the average of the gradients of all pixels, used as a benchmark for comparison. The gradient change intensity expressed as a percentage provides an intuitive view of the rate of change. The summation symbol indicates that the gradients of all pixels in the specified area are cumulatively analyzed.

[0119] S303: Call the transformation kernel matrix, combine the brightness gradient to calculate the neighborhood brightness adjustment value, extract the brightness change trend of the neighborhood of the pixel, identify the brightness change coefficient, and correct the brightness of the image pixels according to the adjustment coefficient, integrate the adjusted image data, and obtain the super-resolution enhanced image matrix.

[0120] The transformation kernel matrix is ​​invoked, and the neighborhood brightness adjustment value is calculated in conjunction with the brightness gradient. The brightness gradient calculation window size is set to 7×7 pixels. The average brightness value of the pixels within the window is calculated, and the brightness gradient change trend is calculated. The judgment criterion for the brightness change trend is set, and 1.4 times the global brightness gradient average is taken as the reference value. Pixels that meet the standard are selected, and the brightness change coefficient is calculated. The brightness change coefficient is defined as the ratio of the brightness gradient change rate to the transformation kernel parameter. The image pixel brightness is corrected according to the brightness change coefficient, and the pixel brightness value is adjusted to keep it within the range of 85% to 120% of the original brightness to avoid over-enhancing or over-suppressing brightness. All adjusted image data are calculated, integrated, and finally, the super-resolution enhanced image matrix is ​​obtained.

[0121] Please see Figure 5 The specific steps of S4 are as follows:

[0122] S401: Call the super-resolution enhanced image matrix, extract the short-time frequency features of the echo signal in the nodule region, calculate the instantaneous frequency value of the pixel, obtain the frequency change trend of the local region, identify the instantaneous frequency peak, and generate the instantaneous frequency peak matrix.

[0123] Using digital signal processing techniques, the instantaneous frequency value of each pixel within the nodule region is calculated. A 5x5 pixel Short-Time Fourier Transform (STFT) window is set, and STFT is performed on each pixel to extract the frequency corresponding to the maximum energy peak in the spectrum as the instantaneous frequency of that point. The instantaneous frequency of each pixel is recorded, and an instantaneous frequency distribution map of the entire nodule region is plotted. A local region is defined as a 10x10 pixel sub-window, and the average frequency value within each local region is calculated. The local frequency variation trend is observed, and by comparing the average frequencies of adjacent sub-windows, peak regions of instantaneous frequencies are identified. These peaks reflect significant frequency changes in the local region. These data are integrated into an instantaneous frequency peak matrix, ultimately forming a matrix containing all peak frequency information of the nodule region.

[0124] S402: Based on the instantaneous frequency peak matrix, calculate the frequency change of the pixel in the spatial position, extract the frequency change difference between pixels, count the rate of change in the local frequency gradient direction, calculate the spatial frequency offset amplitude, analyze the local change trend of the offset distribution, construct the offset distribution matrix according to the frequency change rate, filter the regions with offset values ​​higher than the set threshold, and obtain the offset distribution matrix.

[0125] The gradient operator is used to calculate the frequency change of each pixel in spatial location. This involves calculating the frequency difference between adjacent pixels. The central difference method is used to calculate the frequency changes in the horizontal and vertical directions. These changes reveal the spatial variation pattern of frequency. Statistical analysis is performed on these frequency difference data to calculate the average rate and standard deviation of frequency change in local regions. This provides a basis for analyzing the local variation trend of the offset distribution. A frequency change rate threshold is defined, and all regions exceeding this threshold are marked as significant change regions. These significant change regions may indicate pathological abnormalities. Through these analyses, an offset distribution matrix is ​​constructed, recording the frequency offset amplitude of each pixel, and finally obtaining a matrix that describes the spatial frequency offset characteristics in detail.

[0126] S403: Call the offset distribution matrix, analyze the variation characteristics of the spatial frequency gradient, calculate the degree of local frequency drift, filter the frequency offset of abnormal areas, integrate the frequency change data, and obtain the echo frequency offset matrix.

[0127] By analyzing the spatial frequency gradient variation characteristics recorded in the data, a mathematical model is used to calculate the local frequency gradient and identify abnormal regions. These abnormal regions are marked by comparing the difference between the frequency offset and the average offset. The abnormal screening threshold is defined as the average offset plus twice the standard deviation. This screening helps to identify regions whose frequency changes are significantly different from those of the surrounding structures. This analysis helps to locate possible lesion areas. These frequency offset data are integrated to calculate the average frequency offset of each abnormal region and assess its impact on the overall image. Finally, an echo frequency offset matrix is ​​formed. This matrix provides a new approach to diagnosing diseases through frequency analysis.

[0128] Please see Figure 6 The specific steps of S5 are as follows:

[0129] S501: Call the echo frequency offset matrix, count the spatial distribution density of the abnormal frequency shift region inside the nodule, calculate the proportion of the number of abnormal frequency shift points in the region, calculate the spatial coverage of the abnormal region based on the local distribution, and obtain the abnormal frequency shift spatial density matrix.

[0130] The matrix is ​​traversed pixel by pixel, with a window size of 10×10 pixels. The number of abnormal frequency shift points in each window is counted. An abnormal frequency shift point is defined as a pixel whose offset exceeds the global frequency shift mean plus 2 standard deviations. The proportion of abnormal frequency shift points in each window is calculated. The spatial coverage of the abnormal region is calculated using the local distribution. The abnormal frequency shift points of the entire nodule region are accumulated and the proportion of the abnormal region to the total nodule area is calculated. A spatial coverage threshold is defined and set to 1.5 times the global average proportion of abnormal frequency shift points. Regions with coverage exceeding this threshold are filtered out. The statistical data of all windows are integrated to finally obtain the abnormal frequency shift spatial density matrix.

[0131] S502: Based on the abnormal frequency shift spatial density matrix, calculate the coverage ratio of the abnormal region within the nodule range, statistically analyze the frequency shift value variation range of the covered region, calculate the frequency fluctuation rate of the abnormal region, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient based on the frequency shift mean of the differentiated region, filter the region with stability below the threshold, extract the abnormal frequency shift stability features, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix.

[0132] The calculation window size is set to 15×15 pixels. The proportion of abnormal frequency shift points in each window is counted, and the abnormal coverage ratio of the entire nodule region is calculated. The range of frequency shift value changes in each window is compared, and the maximum and minimum difference of frequency shift values ​​in each window is calculated. A sliding window is used to calculate the local frequency fluctuation rate, which is defined as the mean change of frequency shift values ​​between adjacent windows divided by the window step size. The window step size is set to 5 pixels. The coverage ratio data is used to calculate tissue frequency shift stability. A stability calculation standard is set, and stability is defined as the standard deviation of local frequency shift values ​​divided by the mean. The stability coefficient is calculated, and a stability threshold is set. 0.8 times the global stability mean is used as the screening standard to screen areas with stability below the threshold. Abnormal frequency shift stability features are extracted, and the frequency shift stability gradient is calculated. The gradient data is normalized, and all stability calculation results are integrated to finally obtain the tissue frequency shift stability matrix.

[0133] S503: Call the tissue frequency shift stability matrix, calculate the spatial change of frequency shift trend, extract the frequency shift rate of differential regions, statistically analyze the distribution characteristics of the shift change regions, integrate regional frequency shift change data, and obtain ultrasound imaging analysis steps;

[0134] The spatial variation of frequency shift trend is calculated. The window size is set to 20×20 pixels. The frequency shift stability difference between adjacent windows is calculated. The frequency shift offset rate of the differentiated region is calculated using the central difference method. The frequency shift change trend of different regions is compared. The distribution characteristics of the offset change region are statistically analyzed. An offset change threshold is set. 1.2 times the global offset rate mean is taken as the judgment benchmark. The offset rates of all pixels are arranged and regions with values ​​higher than the threshold are selected. Data of all selected regions are integrated to finally obtain the ultrasound imaging analysis steps.

[0135] Please see Figure 7 A thyroid nodule ultrasound examination system, comprising:

[0136] The resonant mode extraction module acquires ultrasound echo signals from thyroid tissue, analyzes the instantaneous amplitude, phase, and short-time frequency characteristics of the echo signals in the nodule region, calculates the energy distribution and instantaneous frequency difference, and establishes a resonant mode feature matrix.

[0137] The signal enhancement and control module calls the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping based on the gain adjustment coefficient, statistically analyzes the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor based on the signal cross-correlation coefficient, performs dynamic weight suppression to adjust the signal strength, and obtains the enhanced echo signal matrix.

[0138] The image gradient optimization module calls the enhanced echo signal matrix to calculate the gray-level gradient of the ultrasound image in the nodule region, constructs the gradient distribution matrix, calls the gradient change rate to calculate the transform kernel matrix, adjusts the pixel contrast based on the transform kernel matrix, calculates the neighborhood brightness adjustment value, and obtains the super-resolution enhanced image matrix.

[0139] The frequency offset detection module calls the super-resolution enhanced image matrix to extract the short-time frequency characteristics of the echo signal in the nodule region, calculates the instantaneous frequency peak, constructs an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to filter abnormal areas, and obtains the echo frequency offset matrix.

[0140] The spatial trend analysis module calls the echo frequency offset matrix to statistically analyze the spatial distribution density of abnormal frequency shift regions within the nodule, calculates the abnormal coverage ratio, calls the frequency shift trend matrix to calculate the offset change rate, and obtains the ultrasound imaging analysis steps.

[0141] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for ultrasound examination of thyroid nodules, characterized in that, Includes the following steps: S1: Obtain ultrasound echo signals from thyroid tissue, analyze the instantaneous characteristics of the nodule region signal, calculate the energy distribution and instantaneous frequency difference, screen stable resonance modes, and establish a resonance mode feature matrix; S2: Call the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculate the gain adjustment coefficient based on the signal peak curve, perform nonlinear mapping based on the gain adjustment coefficient, call the cross-correlation matrix to filter the correlated signals, perform dynamic weight suppression to adjust the signal strength, and obtain the enhanced echo signal matrix; S3: Call the enhanced echo signal matrix, calculate the gray-level gradient of the nodule region, adjust the pixel contrast, and obtain the super-resolution enhanced image matrix; S4: Call the super-resolution enhanced image matrix, extract the short-time frequency features of the nodule region, calculate the frequency change, and obtain the echo frequency offset matrix; S5: Call the echo frequency offset matrix to calculate the abnormal coverage ratio and frequency shift stability, and obtain the ultrasound imaging analysis steps; The specific steps for obtaining the ultrasound imaging analysis steps, including calling the echo frequency offset matrix, calculating the abnormal coverage ratio and frequency shift stability, are as follows: S501: Call the echo frequency offset matrix, count the spatial distribution density of the abnormal frequency shift region inside the nodule, calculate the proportion of the number of abnormal frequency shift points in the region, calculate the spatial coverage of the abnormal region based on the local distribution, and obtain the abnormal frequency shift spatial density matrix. S502: Based on the abnormal frequency shift spatial density matrix, calculate the coverage ratio of the abnormal region within the nodule range, statistically analyze the frequency shift value variation range of the covered region, calculate the frequency fluctuation rate of the abnormal region, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient based on the frequency shift mean of the differentiated region, filter regions with stability below the threshold, extract abnormal frequency shift stability features, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix. S503: Call the tissue frequency shift stability matrix, calculate the spatial change of frequency shift trend, extract the frequency shift rate of differential regions, statistically analyze the distribution characteristics of the shift change regions, integrate regional frequency shift change data, and obtain ultrasound imaging analysis steps.

2. The ultrasound examination method for thyroid nodules according to claim 1, characterized in that, The resonant mode feature matrix includes modal energy characteristics, resonant frequency distribution, and instantaneous amplitude parameters. The enhanced echo signal matrix includes signal gain adjustment coefficient, noise suppression weight, and target signal energy ratio. The super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness equalization adjustment coefficients. The echo frequency offset matrix includes spatial frequency shift amplitude, local frequency anomaly index, and main frequency peak change rate. The ultrasound imaging analysis steps include abnormal area coverage, frequency offset stability, and nodule tissue frequency shift characteristics.

3. The ultrasound examination method for thyroid nodules according to claim 1, characterized in that, The specific steps for acquiring ultrasound echo signals from thyroid tissue, analyzing the instantaneous characteristics of the nodule region signal, calculating the energy distribution and instantaneous frequency difference, screening stable resonance modes, and establishing the resonance mode feature matrix are as follows: S101: Acquire ultrasound echo signals of thyroid tissue, analyze the instantaneous amplitude, phase and short-time frequency of the signal, calculate the amplitude change rate, phase shift and short-time frequency change within the time window, and obtain the instantaneous signal feature matrix; S102: Based on the instantaneous signal feature matrix, calculate the energy distribution of the frequency band signal, obtain the signal energy of the differentiated frequency bands, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency bands, calculate the frequency change rate, analyze the frequency difference distribution between frequency bands, determine the trend of frequency difference in the time dimension, and obtain the instantaneous frequency difference matrix. S103: Call the instantaneous frequency difference matrix, filter the resonant modes with stable frequency differences, extract the signal modes that meet the threshold requirements, integrate the corresponding amplitude, phase and short-time frequency data, and establish the resonant mode feature matrix.

4. The ultrasound examination method for thyroid nodules according to claim 1, characterized in that, The specific steps for obtaining the enhanced echo signal matrix are as follows: The resonant mode feature matrix is ​​invoked to calculate the modal energy density of the echo signal in the nodal region; the gain adjustment coefficient is calculated based on the signal peak curve; nonlinear mapping is performed according to the gain adjustment coefficient; the cross-correlation matrix is ​​invoked to filter correlated signals; and dynamic weight suppression is performed to adjust the signal strength. S201: Call the resonant mode feature matrix to analyze the modal energy density of the echo signal in the nodule region, extract the energy value of the frequency band signal, calculate the average energy value per unit area, and statistically analyze the energy fluctuation range to obtain the modal energy peak value and corresponding frequency, and obtain the modal energy density matrix. S202: Based on the modal energy density matrix, analyze the gain adjustment coefficient of the signal peak curve, call the energy change rate analysis of the peak region to analyze the gain adjustment factor, establish a nonlinear mapping function according to the energy gradient, adjust the signal intensity, calculate the cross-correlation value of the background organization modal signal, construct the cross-correlation matrix between signals, filter signals that meet the correlation threshold, adjust the signal mapping parameters according to the correlation coefficient, filter the background signals that meet the constraints, and obtain the signal cross-correlation matrix. S203: Call the aforementioned signal cross-correlation matrix, calculate the signal cross-correlation coefficient, calculate the dynamic suppression factor based on the correlation coefficient, adjust the signal strength, perform dynamic weighted suppression, integrate the adjusted signal matrix, and obtain the enhanced echo signal matrix.

5. The ultrasound examination method for thyroid nodules according to claim 4, characterized in that, The formula for calculating the average energy is as follows: ； in, The mean modal energy of a unit region. The number of sampling points within a unit area. Representing the The real part of the signal component at each sampling point Representing the The imaginary part of the signal at each sampling point Representing the The instantaneous frequency of each sampling point It represents the average instantaneous frequency of all sampling points within a unit area.

6. The ultrasound examination method for thyroid nodules according to claim 1, characterized in that, The specific steps for calling the enhanced echo signal matrix, calculating the gray-level gradient of the nodule region, adjusting pixel contrast, and obtaining the super-resolution enhanced image matrix are as follows: S301: Call the enhanced echo signal matrix to calculate the gray-level gradient of the ultrasound image of the nodule region, extract the gray-level change rate of each pixel in the image, calculate the gradient change amplitude between pixels based on the gradient distribution, construct the gray-level gradient distribution matrix, and obtain the gradient distribution matrix. S302: Based on the gradient distribution matrix, analyze the gradient change rate, statistically analyze the gradient change trend of pixels in the image area, determine the transformation kernel parameters according to the gradient change rate, call the transformation kernel parameters to adjust the pixel contrast, calculate the contrast change value of the pixel, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix. S303: Call the transformation kernel matrix, calculate the neighborhood brightness adjustment value in combination with the brightness gradient, extract the neighborhood brightness change trend of the pixel, identify the brightness change coefficient, correct the image pixel brightness according to the adjustment coefficient, integrate the adjusted image data, and obtain the super-resolution enhanced image matrix.

7. The ultrasound examination method for thyroid nodules according to claim 6, characterized in that, The gradient change trend of pixels within the statistical image area is expressed by the formula: ； The transformation kernel parameters are determined based on the gradient change rate. The transformation kernel parameters are then used to adjust the pixel contrast. The contrast change value of each pixel is calculated. The image contrast distribution is adjusted based on the change value to obtain the transformation kernel matrix. in, Represents the percentage of the average gradient rate of change. Represents the total number of pixels within the calculation area. Representing the Gradient values ​​of pixels, The summation sign represents the average gradient values ​​of all pixels within the region. This indicates that the gradient changes of all pixels are accumulated.

8. The ultrasound examination method for thyroid nodules according to claim 1, characterized in that, The specific steps for using the super-resolution enhanced image matrix to extract short-time frequency features of the nodule region, calculate frequency changes, and obtain the echo frequency offset matrix are as follows: S401: Call the super-resolution enhanced image matrix, extract the short-time frequency features of the echo signal in the nodule region, calculate the instantaneous frequency value of the pixel, obtain the frequency change trend of the local region, identify the instantaneous frequency peak, and generate an instantaneous frequency peak matrix. S402: Based on the instantaneous frequency peak matrix, calculate the frequency change of the pixel in the spatial position, extract the frequency change difference between pixels, count the rate of change in the local frequency gradient direction, calculate the spatial frequency offset amplitude value, analyze the local change trend of the offset distribution, construct the offset distribution matrix according to the frequency change rate, filter the regions with offset values ​​higher than a set threshold, and obtain the offset distribution matrix. S403: Call the offset distribution matrix to analyze the change characteristics of the spatial frequency gradient, calculate the degree of local frequency drift, filter the frequency offset of abnormal areas, integrate the frequency change data, and obtain the echo frequency offset matrix.

9. A thyroid nodule ultrasound examination system, characterized in that, A method for ultrasound examination of thyroid nodules according to any one of claims 1-8, wherein the system comprises: The resonant mode extraction module acquires ultrasound echo signals from thyroid tissue, analyzes the instantaneous amplitude, phase, and short-time frequency characteristics of the echo signals in the nodule region, calculates the energy distribution and instantaneous frequency difference, and establishes a resonant mode feature matrix. The signal enhancement and control module calls the resonant mode feature matrix to calculate the modal energy density of the echo signal in the nodule region, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping based on the gain adjustment coefficient, statistically analyzes the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor based on the signal cross-correlation coefficient, performs dynamic weight suppression to adjust the signal strength, and obtains the enhanced echo signal matrix. The image gradient optimization module calls the enhanced echo signal matrix to calculate the gray-level gradient of the ultrasound image in the nodule region, constructs a gradient distribution matrix, calls the gradient change rate to calculate the transform kernel matrix, adjusts the pixel contrast based on the transform kernel matrix, calculates the neighborhood brightness adjustment value, and obtains the super-resolution enhanced image matrix. The frequency offset detection module calls the super-resolution enhanced image matrix to extract the short-time frequency characteristics of the echo signal in the nodule region, calculates the instantaneous frequency peak, constructs an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to filter abnormal regions, and obtains the echo frequency offset matrix. The spatial trend analysis module calls the echo frequency offset matrix to statistically analyze the spatial distribution density of abnormal frequency shift regions inside the nodule, calculates the abnormal coverage ratio, calls the frequency shift trend matrix to calculate the offset change rate, and obtains the ultrasound imaging analysis steps. The specific steps for obtaining the ultrasound imaging analysis steps, including calling the echo frequency offset matrix, calculating the abnormal coverage ratio and frequency shift stability, are as follows: S501: Call the echo frequency offset matrix, count the spatial distribution density of the abnormal frequency shift region inside the nodule, calculate the proportion of the number of abnormal frequency shift points in the region, calculate the spatial coverage of the abnormal region based on the local distribution, and obtain the abnormal frequency shift spatial density matrix. S502: Based on the abnormal frequency shift spatial density matrix, calculate the coverage ratio of the abnormal region within the nodule range, statistically analyze the frequency shift value variation range of the covered region, calculate the frequency fluctuation rate of the abnormal region, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient based on the frequency shift mean of the differentiated region, filter regions with stability below the threshold, extract abnormal frequency shift stability features, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix. S503: Call the tissue frequency shift stability matrix, calculate the spatial change of frequency shift trend, extract the frequency shift rate of differential regions, statistically analyze the distribution characteristics of the shift change regions, integrate regional frequency shift change data, and obtain ultrasound imaging analysis steps.

Images