Thyroid nodule ultrasonic examination method and system

Abstract

The invention relates to the technical field of ultrasonic imaging analysis, in particular to a thyroid nodule ultrasonic examination method and system.The thyroid nodule ultrasonic examination method comprises the following steps that thyroid tissue ultrasonic echo signals are obtained, instantaneous characteristics of nodule area signals are analyzed, energy distribution and instantaneous frequency difference values are calculated, stable resonance modes are screened, and a resonance mode characteristic matrix is established. According to the method, the decomposition precision of echo data and the accurate recognition capability of a target signal are improved by optimizing instantaneous feature analysis of an ultrasonic signal, background noise interference is remarkably reduced by combining energy distribution and frequency difference screening, image enhancement is optimized through a gradient change rate and a transformation kernel matrix, the definition of image details is improved, and the recognition accuracy of the target signal is improved. Short-time frequency feature extraction and frequency deviation measurement enhance the recognition capability of lesion tissues, and meanwhile, through analysis of frequency shift stability parameters and the deviation change rate, the accuracy of automatic judgment of nodule properties is improved, so that the thyroid nodule ultrasonic examination is more scientific and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic imaging analysis, and in particular to a method and system for ultrasonic examination of thyroid nodules. Background Art

[0002] The field of ultrasonic imaging and analysis technology includes ultrasonic imaging and data processing methods in medical imaging. The core content involves the acquisition, transmission, imaging and analysis of ultrasonic signals, including the design and application of ultrasonic transducers, the propagation characteristics of ultrasonic signals, the reception and processing of echo signals, image reconstruction and analysis methods, etc. In medical diagnosis, ultrasonic imaging is widely used in the visualization detection of organs and tissues, and real-time non-invasive examination is achieved through ultrasonic signals of different frequencies and modes. Systematic research in this technical field covers the construction of ultrasonic imaging equipment, ultrasonic emission and reception modes, tissue echo signal feature analysis, image enhancement and denoising processing, automatic identification and classification of target areas, etc., especially for the acoustic characteristics of different tissue structures, using specific parameters to optimize imaging quality, and combining computer-aided technology for automated data analysis and identification.

[0003] Among them, the ultrasound examination method of thyroid nodules refers to the inspection method that uses ultrasound imaging technology to identify and classify the structure, morphology, boundaries, internal echo characteristics and blood flow status of thyroid nodules. The echo data of thyroid tissue is obtained through high-frequency ultrasound signals, and the amplitude, phase and spectrum characteristics of the echo signals are extracted and analyzed to identify the echo type and tissue composition of the nodules. It includes the analysis of the boundary characteristics of the nodules to determine their contour integrity, morphological regularity and relationship with surrounding tissues. At the same time, combined with blood flow signal analysis technology, Doppler ultrasound is used to detect the distribution of blood vessels inside and outside the nodules. Finally, the imaging characteristics of thyroid nodules are determined through image feature matching and classification standards, providing a basis for subsequent clinical evaluation.

[0004] Existing technologies mainly rely on the amplitude and phase information of traditional echo signals in signal analysis, and do not fully consider the impact of instantaneous frequency changes, resulting in low signal decomposition accuracy and susceptibility to interference from tissue scattering effects. During the signal enhancement process, the background noise suppression strategy is relatively simple and fails to be dynamically adjusted according to different tissue characteristics, which may lead to insufficient signal enhancement or target signal loss. Image optimization relies on conventional contrast enhancement and edge sharpening techniques, which makes it difficult to effectively distinguish the boundaries of low-contrast nodules, affecting the morphological analysis of nodules. Frequency offset measurement fails to fully incorporate spatial position information, which may lead to the omission of local abnormal frequency shift signals and reduce the accuracy of diseased tissue identification. The lack of analysis of the spatial distribution and stability parameters of abnormal areas makes the determination of nodule properties rely on subjective experience, increasing diagnostic uncertainty. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a method and system for ultrasonic examination of thyroid nodules.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: a method for ultrasonic examination of thyroid nodules, comprising the following steps:

[0007] S1: Acquire the ultrasonic echo signal of thyroid tissue, analyze the instantaneous characteristics of the signal in the nodule area, calculate the energy distribution and instantaneous frequency difference, screen the stable resonance mode, and establish the resonance mode characteristic matrix;

[0008] S2: calling the resonance modal characteristic matrix, calculating the modal energy density of the echo signal in the nodule area, calculating the gain adjustment coefficient based on the signal peak curve, performing nonlinear mapping according to the gain adjustment coefficient, calling the mutual correlation matrix to screen the correlation signal, performing dynamic weight suppression to adjust the signal strength, and obtaining an enhanced echo signal matrix;

[0009] S3: calling the enhanced echo signal matrix, calculating the grayscale gradient of the nodule area, adjusting the pixel contrast, and obtaining a super-resolution enhanced image matrix;

[0010] S4: calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the nodule area, calculating the frequency change, and obtaining the echo frequency offset matrix;

[0011] S5: calling the echo frequency offset matrix, calculating the abnormal coverage ratio and the frequency shift stability, and obtaining the ultrasonic imaging analysis steps.

[0012] As a further solution of the present invention, the resonance modal feature matrix includes modal energy characteristics, resonance frequency distribution, and instantaneous amplitude parameters; the enhanced echo signal matrix includes signal gain adjustment coefficients, noise suppression weights, and target signal energy ratios; the super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness balance 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 step includes abnormal area coverage, frequency offset stability, and nodule tissue frequency shift characteristics.

[0013] As a further solution of the present invention, the specific steps of obtaining the ultrasonic echo signal of the thyroid tissue, analyzing the instantaneous characteristics of the nodule area signal, calculating the energy distribution and the instantaneous frequency difference, screening the stable resonance mode, and establishing the resonance mode characteristic matrix are as follows:

[0014] S101: Acquire thyroid tissue ultrasonic echo signals, analyze the instantaneous amplitude, phase and short-time frequency of the signals, calculate the amplitude change rate, phase offset and short-time frequency variation within the time window, and obtain an 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 band, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency band, calculate the frequency change rate, analyze the frequency difference distribution between the frequency bands, determine the change trend of the frequency difference in the time dimension, and obtain the instantaneous frequency difference matrix;

[0016] S103: calling the instantaneous frequency difference matrix, screening the resonance modes with stable frequency differences, extracting the signal modes that meet the threshold requirements, integrating the corresponding amplitude, phase and short-time frequency data, and establishing the resonance mode feature matrix.

[0017] As a further solution of the present invention, the resonance modal feature matrix is ​​called, the modal energy density of the echo signal in the nodule area is calculated, the gain adjustment coefficient is calculated based on the signal peak curve, nonlinear mapping is performed according to the gain adjustment coefficient, the mutual correlation matrix is ​​called to screen the correlation signal, and the dynamic weight suppression is performed to adjust the signal strength. The specific steps of obtaining the enhanced echo signal matrix are:

[0018] S201: calling the resonance modal characteristic matrix, analyzing the modal energy density of the echo signal in the nodule area, extracting the energy value of the frequency band signal, calculating the energy mean of the unit area, and counting the energy fluctuation range, obtaining the modal energy peak value and the corresponding frequency, and obtaining 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 of the peak area 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 tissue modal signal, construct a cross-correlation matrix between signals, screen the signal that meets the correlation threshold, adjust the signal mapping parameters according to the correlation coefficient, screen the background signal that meets the constraint, and obtain the signal cross-correlation matrix;

[0020] S203: calling the signal correlation matrix, calculating the signal correlation coefficient, calculating the dynamic suppression factor according to the correlation coefficient, adjusting the signal strength, performing dynamic weight suppression, integrating the adjusted signal matrix, and obtaining an enhanced echo signal matrix.

[0021] As a further solution of the present invention, the energy mean calculation formula is specifically:

[0022]

[0023] Among them, E avg represents the mean modal energy per unit area, N represents the number of sampling points per unit area, and A i represents the real signal component of the i-th sampling point, B irepresents the imaginary signal component of the i-th sampling point, C i represents the instantaneous frequency of the i-th sampling point, Represents the instantaneous frequency mean of all sampling points in the unit area.

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

[0025] S301: calling the enhanced echo signal matrix, calculating the grayscale gradient of the ultrasound image of the nodule area, extracting the grayscale change rate of each pixel in the image, calculating the gradient change amplitude between the pixels based on the gradient distribution, constructing the grayscale gradient distribution matrix, and obtaining the gradient distribution matrix;

[0026] S302: Based on the gradient distribution matrix, analyze the gradient change rate, count the gradient change trend of the pixel points 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 points, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix;

[0027] S303: calling the transformation kernel matrix, calculating the neighborhood brightness adjustment value in combination with the brightness gradient, extracting the neighborhood brightness change trend of the pixel point, identifying the brightness change coefficient, and correcting the image pixel brightness according to the adjustment coefficient, integrating the adjusted image data, and obtaining the super-resolution enhanced image matrix.

[0028] As a further solution of the present invention, the gradient change trend of the pixels in the statistical image area is calculated using the formula:

[0029]

[0030] 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 point, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix;

[0031] Among them, ΔK represents the average gradient change rate percentage, M represents the total number of pixels in the calculation area, and G k Represents the gradient value of the kth pixel, G avg Represents the average value of all pixel gradient values ​​in the area, sum symbol Indicates the accumulation of gradient changes of all pixels.

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

[0033] S401: calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the echo signal in the nodule area, calculating the instantaneous frequency value of the pixel point, obtaining the frequency change trend of the local area, identifying the instantaneous frequency peak, and generating the instantaneous frequency peak matrix;

[0034] S402: Based on the instantaneous frequency peak matrix, the frequency change of the pixel point in the spatial position is calculated, the frequency change difference between the pixels is extracted, the change rate in the local frequency gradient direction is counted, the spatial frequency offset amplitude value is calculated, the local change trend of the offset distribution is analyzed, and the offset distribution matrix is ​​constructed according to the frequency change rate. The area where the offset is higher than the set threshold is screened to obtain the offset distribution matrix;

[0035] S403: calling the offset distribution matrix, analyzing the variation characteristics of the spatial frequency gradient, calculating the degree of local frequency drift, screening the frequency offset amount in the abnormal area, integrating the frequency variation data, and obtaining the echo frequency offset matrix.

[0036] As a further solution of the present invention, the specific steps of calling the echo frequency offset matrix, calculating the abnormal coverage ratio and the frequency shift stability, and obtaining the ultrasonic imaging analysis step are:

[0037] S501: calling the echo frequency shift matrix, counting the spatial distribution density of the abnormal frequency shift area inside the nodule, calculating the number ratio of the abnormal frequency shift points in the area, calculating the spatial coverage of the abnormal area according to the local distribution, and obtaining 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 area within the nodule range, count the frequency shift value variation range of the coverage area, calculate the frequency fluctuation rate of the abnormal area, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient according to the frequency shift mean of the differentiated area, screen the area with stability lower than the threshold, extract the abnormal frequency shift stability feature, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix;

[0039] S503: calling the tissue frequency shift stability matrix, calculating the spatial variation of the frequency shift trend, extracting the frequency shift offset rate of the differentiated area, statistically analyzing the distribution characteristics of the offset change area, integrating the regional frequency shift change data, and obtaining the ultrasound imaging analysis step.

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

[0041] The resonance mode extraction module obtains the ultrasonic echo signal of the thyroid tissue, analyzes the instantaneous amplitude, phase and short-time frequency characteristics of the echo signal in the nodule area, calculates the energy distribution and instantaneous frequency difference, and establishes the resonance mode feature matrix;

[0042] The signal enhancement control module calls the resonance modal feature matrix, calculates the modal energy density of the echo signal in the nodule area, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping according to the gain adjustment coefficient, counts the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor according to the signal cross-correlation coefficient, performs dynamic weight suppression to adjust the signal strength, and obtains an enhanced echo signal matrix;

[0043] The image gradient optimization module calls the enhanced echo signal matrix, calculates the grayscale gradient of the ultrasound image of the nodule area, constructs a gradient distribution matrix, calls the gradient change rate to calculate the transformation kernel matrix, adjusts the pixel contrast according to the transformation 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 imaging matrix, extracts the short-time frequency characteristics of the echo signal in the nodule area, calculates the instantaneous frequency peak, constructs an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to screen the abnormal area, and obtains the echo frequency offset matrix;

[0045] The spatial trend analysis module calls the echo frequency offset matrix, counts the spatial distribution density of the abnormal frequency shift area 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.

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

[0047] In the present invention, by optimizing the instantaneous characteristic analysis of ultrasonic signals, the decomposition accuracy of echo data and the accurate recognition ability 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 offset measurement enhance the recognition ability of diseased tissue. At the same time, through the analysis of frequency shift stability parameters and offset change rate, the accuracy of automatic judgment of nodule properties is improved, making thyroid nodule ultrasonic examination more scientific and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

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

[0050] Figure 2is a flow chart of the steps of S1 of the present invention;

[0051] Figure 3 is a flow chart of the steps of S2 of the present invention;

[0052] Figure 4 is a flow chart of the steps of S3 of the present invention;

[0053] Figure 5 is a flow chart of the steps of S4 of the present invention;

[0054] Figure 6 is a flow chart of the steps of S5 of the present invention;

[0055] Figure 7 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[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 "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0058] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express 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 more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0061] See also Figure 1 , a method for ultrasonic examination of thyroid nodules, comprising the following steps:

[0062] S1: Acquire the ultrasonic echo signal of thyroid tissue, analyze the instantaneous amplitude, phase and short-time frequency characteristics of the echo signal in the nodule area, calculate the energy distribution and instantaneous frequency difference, call the frequency difference to screen the stable resonance mode, and establish the resonance mode characteristic matrix;

[0063] S2: call the resonance modal feature matrix, calculate the modal energy density of the echo signal in the nodule area, calculate the gain adjustment coefficient based on the signal peak curve, perform nonlinear mapping according to the gain adjustment coefficient, count the cross-correlation values ​​of the background tissue modal signal, call the cross-correlation matrix to screen the correlation signal, calculate the dynamic suppression factor according to 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 grayscale gradient of the ultrasound image of the nodule area, construct the gradient distribution matrix, call the gradient change rate to calculate the transformation kernel matrix, adjust the pixel contrast according to the transformation kernel matrix, combine the brightness gradient, calculate the neighborhood brightness adjustment value, and obtain the super-resolution enhanced image matrix;

[0065] S4: call the super-resolution enhanced imaging matrix, extract the short-time frequency characteristics of the echo signal in the nodule area, calculate the instantaneous frequency peak, call the peak matrix to calculate the frequency change in the spatial position, build the offset distribution matrix based on the frequency offset amplitude, call the offset distribution matrix to screen the abnormal area, and obtain the echo frequency offset matrix;

[0066] S5: Call the echo frequency offset matrix, count the spatial distribution density of the abnormal frequency shift area inside the nodule, calculate the abnormal coverage ratio, call the coverage ratio matrix to calculate the 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 resonance modal feature matrix includes modal energy characteristics, resonance frequency distribution, and instantaneous amplitude parameters. The enhanced echo signal matrix includes signal gain adjustment coefficients, noise suppression weights, and target signal energy ratios. The super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness balance 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.

[0068] See also Figure 2 , the specific steps of S1 are:

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

[0070] Obtain the ultrasonic echo signal of thyroid tissue, collect data on the echo signal, and set the sampling frequency f s The sampling time is 40 MHz, the sampling time T is 50 ms, and the number of signal data points obtained is N = f s ×T=2×10 6 , bandpass filtering is performed on the signal data, and the upper and lower limit frequencies of the filter are set to 2MHz and 15MHz respectively, only the effective frequency band signal of the thyroid tissue echo is retained, and the Hilbert transform is performed on the filtered signal to extract the instantaneous amplitude, instantaneous phase and short-time frequency of the signal. The instantaneous amplitude A(t) is obtained by calculating the envelope of the signal, and the instantaneous phase φ(t) is obtained by calculating the inverse tangent 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 taking the derivative of the phase with respect to time. The signal is processed in segments using the sliding window method. The window length is set to 512 points and the window sliding step is 128 points. The amplitude change rate is calculated in each window and is defined as dA / dt. The amplitude difference between adjacent sampling points is divided by the time step Δt=1 / f s Obtain and calculate the phase offset, which is defined as the cumulative change of phase Δφ=φ(t+Δt)-φ(t), and calculate the short-time frequency variation, which is defined as df / dt, obtained by dividing the short-time frequency difference between adjacent moments by the time step, and arrange these parameters in chronological order to form an 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 band, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency band, calculate the frequency change rate, analyze the frequency difference distribution between the frequency bands, determine the change 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, and the signal is subjected to short-time Fourier transform (STFT). The window function is set to Hanning window, the window length is 512 points, and the step length is 128 points. The time-frequency spectrum energy distribution matrix E(f, t) is calculated, and the frequency bands with higher energy are selected, and the energy threshold E is set. th The mean of the energy distribution plus twice the standard deviation, that is, E th =μ E +2σ E , select the frequency band that meets the threshold as the differentiated frequency band, and calculate the instantaneous frequency f of the differentiated frequency band d (t), perform differential operation on the instantaneous frequency and calculate the instantaneous frequency difference Δf=f between adjacent time points d (t+Δt)-f d (t), calculate the rate of change of frequency, defined as df d / dt, the instantaneous frequency difference between adjacent moments is obtained by dividing the time step, analyzing the frequency difference distribution between different frequency bands, and setting the time window T w =10ms, calculate the mean and standard deviation of the frequency difference in the time window, determine the changing trend of the frequency difference in the time dimension, use the linear regression method to fit the changing trend of the frequency difference, and perform statistics on the fitting slope k. If |k| is less than the set threshold k th =0.05Hz / ms, it is determined that the frequency difference change in this frequency band is stable, and finally an instantaneous frequency difference matrix is ​​formed.

[0073] S103: calling the instantaneous frequency difference matrix, screening the resonance mode with stable frequency difference, extracting the signal mode that meets the threshold requirement, integrating the corresponding amplitude, phase and short-time frequency data, and establishing the resonance mode characteristic matrix;

[0074] Screen the resonance modes with stable frequency difference, perform statistical analysis on each frequency band in the matrix, and calculate the standard deviation σ of the frequency difference of each frequency band in the entire signal duration f , set the threshold σ th is the median of the frequency difference, and σ is selected f <σ th The frequency band is taken as the stable resonance mode, the signal mode that meets the threshold requirement is extracted, the amplitude, phase and short-time frequency data of the corresponding frequency band are called from the instantaneous signal feature matrix, and the amplitude threshold A is set. th is 1.5 times the standard deviation of the amplitude mean, that is, A th =μ A +1.5σ A , the screening amplitude is greater than A th signal points, integrate the data that meet the screening criteria, establish the resonance mode feature matrix, and finally form a data set for subsequent analysis.

[0075] See also Figure 3 , the specific steps of S2 are:

[0076] S201: calling the resonance modal characteristic matrix, analyzing the modal energy density of the echo signal in the nodule area, extracting the energy value of the frequency band signal, calculating the energy mean of the unit area, and counting the energy fluctuation range, obtaining the modal energy peak value and the corresponding frequency, and obtaining the modal energy density matrix;

[0077] The specific formula for calculating the mean energy is:

[0078]

[0079] Among them, E avg represents the mean modal energy per unit area, N represents the number of sampling points per unit area, and A irepresents the real signal component of the i-th sampling point, B i represents the imaginary signal component of the i-th sampling point, C i represents the instantaneous frequency of the i-th sampling point, Represents the instantaneous frequency mean of all sampling points in the unit area.

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

[0081] Calculate the complex signal amplitude at each sampling point Among them A i and B i are the real and imaginary components of the ith sampling point, respectively. The calculation of this step is based on the complex representation of the signal, which is a commonly used method in electrical signal processing to obtain the instantaneous amplitude of the signal.

[0082] Calculate the frequency deviation at each sampling point Among them C i is the instantaneous frequency of the ith sampling point, and is the mean of the instantaneous frequencies of all sampling points in the entire sampling area. This step provides the degree of change in the frequency of each point. Large frequency deviations may indicate areas with more concentrated energy or more significant changes.

[0083] Sum and multiply the above two results, and then divide by the total number of sampling points N to obtain the energy mean E of the entire area avg .

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

[0085] Assume that there are N = 4 sampling points in a unit area, 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=5.5Hz

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

[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 of each point and substitute into the formula:

[0093]

[0094]

[0095] Finally, calculate E avg :

[0096]

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

[0098] S202: Based on the modal energy density matrix, analyze the gain adjustment coefficient of the signal peak curve, call the energy change rate of the peak area 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 tissue modal signal, construct a signal cross-correlation matrix, screen the signal that meets the correlation threshold, adjust the signal mapping parameters according to the correlation coefficient, screen the background signal that meets the constraint, and obtain the signal cross-correlation matrix;

[0099] Based on the modal energy density matrix, the energy density peak curve is analyzed, the gain adjustment coefficient is set, the energy change rate of the peak area is selected, the energy increment per unit time is calculated, the gain adjustment factor is defined as the ratio of the energy increment to the standard energy, the energy gradient is used to calculate the nonlinear mapping function, the mapping parameter range is set, the energy gain adjustment range is defined between 0.8 and 1.2, the energy change gradient is discretely segmented, and the length of each segment is set to 5% of the maximum energy change of the original signal. The corresponding gain parameters are calculated according to the gradients of different segments, the signal strength is adjusted, the modal signal of the background tissue is cross-correlated, the reference background signal area is set, and the background is calculated. The cross-correlation value of the signal and the nodule signal was used to obtain the mutual correlation matrix between the signals, set the correlation coefficient threshold, take 1.1 times the mean value of the mutual correlation matrix as the screening standard, screen the signals that meet the threshold range, adjust the signal mapping parameters according to the correlation coefficient, set the correlation stratification standard, divide the signals into three groups from high to low according to the correlation coefficient, adjust the mapping parameter range respectively, make the gain adjustment factor of the high correlation coefficient signal between 0.9 and 1.1, the adjustment factor of the medium correlation coefficient signal between 0.85 and 1.15, and the adjustment factor of the low correlation coefficient signal between 0.8 and 1.2, normalize the adjusted signals to form a signal mutual correlation matrix.

[0100] S203: calling the signal cross-correlation matrix, calculating the signal cross-correlation coefficient, calculating the dynamic suppression factor according to the correlation coefficient, adjusting the signal strength, performing dynamic weight suppression, integrating the adjusted signal matrix, and obtaining an enhanced echo signal matrix;

[0101] Call the signal cross-correlation matrix, calculate the signal data in the matrix, calculate the cross-correlation coefficient between each pair of signals, take the mean of the cross-correlation coefficients between all signals as the reference value, set the dynamic suppression factor, compare the cross-correlation coefficient of each signal with the mean, if it is higher than the mean, calculate its relative deviation, set the suppression factor to 0.7 times the deviation, if it is lower than the mean, set the suppression factor to 1.2 times the deviation, adjust the signal strength, multiply each signal data point by its corresponding suppression factor, perform dynamic weight suppression, set the weight adjustment range, keep the suppressed signal amplitude between 70% and 120% of the original signal amplitude, avoid excessive attenuation or enhancement of the signal, calculate all adjusted signal data, integrate them to form an enhanced echo signal matrix.

[0102] See also Figure 4 , the specific steps of S3 are:

[0103] S301: calling the enhanced echo signal matrix, calculating the grayscale gradient of the ultrasound image of the nodule area, extracting the grayscale change rate of each pixel in the image, calculating the gradient change amplitude between the pixels based on the gradient distribution, constructing the grayscale gradient distribution matrix, and obtaining the gradient distribution matrix;

[0104] Call the enhanced echo signal matrix, set the image resolution to 512×512 pixels, calculate the image data pixel by pixel, extract the grayscale value of each pixel, define the grayscale gradient as the grayscale value difference calculation result of adjacent pixels, use Sobel operator for edge detection, calculate the gradient components in the horizontal and vertical directions, set the window size to 3×3 pixels, calculate the grayscale change rate of the central pixel and the surrounding pixels, define the gradient change rate as the gradient change amplitude of each pixel divided by the time step, set the time step to the image frame interval of 0.02 seconds, count the distribution of the gradient change rate in the entire image area, normalize the gradient value, map the gradient amplitude to the interval of 0 to 1, compare the gradient changes between different pixels, set the grayscale gradient distribution threshold, take 1.3 times the global grayscale gradient mean as the standard, screen out the area exceeding the threshold, arrange the gradient data of all pixels to form a grayscale gradient distribution matrix, and finally obtain the gradient distribution matrix.

[0105] S302: Based on the gradient distribution matrix, analyze the gradient change rate, count the gradient change trend of the pixel points 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 points, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix;

[0106] The gradient change trend of pixels in the statistical image area is calculated using the formula:

[0107]

[0108] 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 point, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix;

[0109] Among them, ΔK represents the average gradient change rate percentage, M represents the total number of pixels in the calculation area, and G k Represents the gradient value of the kth pixel, G avg Represents the average value of all pixel gradient values ​​in the area, sum symbol Indicates the accumulation of gradient changes of all pixels;

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

[0111]

[0112] Then, for each pixel, the gradient G k Calculate its average gradient G avg The deviation is calculated with respect to G avg Converted to percentage. This method effectively describes the intensity and direction of pixel gradient changes.

[0113] Set up a specific example: Assume that there are M = 5 pixels in a calculation area, 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 formula to calculate the gradient change rate of each pixel:

[0117]

[0118] The result shows that within a given area, the average gradient change rate is 11.72%, which reflects the average intensity of the gradient change within the area; parameter introduction: ΔK represents the average gradient change rate, which is used to measure the average intensity of pixel gradient changes, M is the total number of pixels considered, which is used to determine the breadth of the calculation, G k is the gradient value of a single pixel, obtained directly from the image processing algorithm, G avgIt is the average value of all pixel gradients and is used as a comparison benchmark. The intensity of gradient change expressed as a percentage provides an intuitive view of the rate of change. The sum symbol represents the cumulative analysis of the gradients of all pixels in the specified area.

[0119] S303: calling the transformation kernel matrix, calculating the neighborhood brightness adjustment value in combination with the brightness gradient, extracting the neighborhood brightness change trend of the pixel point, identifying the brightness change coefficient, and correcting the image pixel brightness according to the adjustment coefficient, integrating the adjusted image data, and obtaining the super-resolution enhanced image matrix;

[0120] The transformation kernel matrix is ​​called, and the neighborhood brightness adjustment value is calculated in combination with the brightness gradient. The brightness gradient calculation window size is set to 7×7 pixels. The brightness values ​​of the pixels in the window are averaged, and the brightness gradient change trend is calculated. The judgment standard of the brightness change trend is set, and 1.4 times the global brightness gradient mean is taken as the reference value. The pixels that meet the standard are screened out, 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 85% to 120% of the original brightness to avoid over-enhancement or over-suppression of brightness. All adjusted image data are calculated and integrated to finally obtain the super-resolution enhanced image matrix.

[0121] See also Figure 5 , the specific steps of S4 are:

[0122] S401: calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the echo signal in the nodule area, calculating the instantaneous frequency value of the pixel point, obtaining the frequency change trend of the local area, identifying the instantaneous frequency peak, and generating the instantaneous frequency peak matrix;

[0123] Through digital signal processing technology, the instantaneous frequency value of each pixel in the nodule area is calculated, the short-time Fourier transform (STFT) window size is set to 5x5 pixels, STFT is performed on each pixel, the frequency corresponding to the maximum energy peak in the spectrum is extracted as the instantaneous frequency of the point, the instantaneous frequency of each pixel is recorded, and the instantaneous frequency distribution diagram of the entire nodule area is drawn. The local area is defined as a 10x10 pixel sub-window, the frequency average value in each local area is calculated, and the local change trend of the frequency is observed. By comparing the average frequencies of adjacent sub-windows, the peak areas of the instantaneous frequency are identified. These peaks reflect the significant frequency changes in the local area. These data are integrated into an instantaneous frequency peak matrix, and finally a matrix containing all peak frequency information of the nodule area is formed.

[0124] S402: Based on the instantaneous frequency peak matrix, the frequency change of the pixel point in the spatial position is calculated, the frequency change difference between the pixels is extracted, the change rate in the local frequency gradient direction is counted, the spatial frequency offset amplitude value is calculated, the local change trend of the offset distribution is analyzed, and the offset distribution matrix is ​​constructed according to the frequency change rate. The area where the offset is higher than the set threshold is screened to obtain the offset distribution matrix;

[0125] The gradient operator is used to calculate the frequency change of each pixel in space, which 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 the frequency. These frequency difference data are statistically analyzed to calculate the average rate and standard deviation of frequency change in the local area, which provides a basis for analyzing the local change trend of the offset distribution. A frequency change rate threshold is defined, and all areas exceeding the threshold are marked as significant change areas. These significant change areas may indicate pathological abnormal changes. Through these analyses, an offset distribution matrix is ​​constructed to record the frequency offset amplitude of each pixel, and finally a matrix that describes the spatial frequency offset characteristics in detail is obtained.

[0126] S403: calling the offset distribution matrix, analyzing the change characteristics of the spatial frequency gradient, calculating the degree of local frequency drift, screening the frequency offset of the abnormal area, integrating the frequency change data, and obtaining the echo frequency offset matrix;

[0127] The recorded spatial frequency gradient change characteristics are analyzed, the local frequency gradient is calculated using a mathematical model, and abnormal areas are identified. These abnormal areas are marked by comparing the difference between the frequency offset and the average offset. The abnormal screening threshold is defined as the offset average plus two standard deviations. Such screening helps identify areas where the frequency changes are significantly different from the surrounding structures. This analysis helps to locate possible lesion areas. These frequency offset data are integrated, the average frequency offset of each abnormal area is calculated, and its impact on the overall image is evaluated, ultimately forming an echo frequency offset matrix. This matrix provides a new way to diagnose diseases through frequency analysis.

[0128] See also Figure 6 , the specific steps of S5 are:

[0129] S501: calling the echo frequency shift matrix, counting the spatial distribution density of the abnormal frequency shift area inside the nodule, calculating the number of abnormal frequency shift points in the area, calculating the spatial coverage of the abnormal area according to the local distribution, and obtaining the abnormal frequency shift spatial density matrix;

[0130] The matrix was traversed at the pixel level, and the window size was set to 10×10 pixels. The number of abnormal frequency shift points in each window was counted, and the abnormal frequency shift points were defined as pixels whose offset exceeded the global frequency shift mean plus 2 times the standard deviation. The proportion of abnormal frequency shift points in each window was calculated, and the spatial coverage of the abnormal area was calculated using the local distribution. The abnormal frequency shift points in the entire nodule area were accumulated, and the proportion of the abnormal area to the total nodule area was calculated. The spatial coverage threshold was defined and set to 1.5 times the global average of the abnormal frequency shift point ratio. The areas with coverage exceeding the threshold were screened out, and the statistical data of all windows were 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 area within the nodule range, count the frequency shift value variation range of the coverage area, calculate the frequency fluctuation rate of the abnormal area, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient according to the frequency shift mean of the differentiated area, screen the area with stability below the threshold, extract the abnormal frequency shift stability feature, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix;

[0132] The calculation window size was set to 15×15 pixels, and the proportion of abnormal frequency shift points in each window was counted. The abnormal coverage ratio of the entire nodule area was calculated. The variation range of the frequency shift values ​​in each window was compared, and the maximum and minimum differences of the frequency shift values ​​in each window were calculated. The local frequency fluctuation rate was calculated using a sliding window. The fluctuation rate was defined as the mean change of the frequency shift values ​​between adjacent windows divided by the window step size. The window step size was set to 5 pixels. The coverage ratio data was called to calculate the tissue frequency shift stability. The stability calculation standard was set. The stability was defined as the standard deviation of the local frequency shift value divided by the mean. The stability coefficient was calculated. The stability threshold was set. 0.8 times the global stability mean was taken as the screening standard. The areas with stability below the threshold were screened. The abnormal frequency shift stability characteristics were extracted. The frequency shift stability gradient was calculated. The gradient data was normalized. All stability calculation results were integrated to finally obtain the tissue frequency shift stability matrix.

[0133] S503: calling the tissue frequency shift stability matrix, calculating the spatial variation of the frequency shift trend, extracting the frequency shift rate of the differentiated area, statistically analyzing the distribution characteristics of the frequency shift change area, integrating the regional frequency shift change data, and obtaining the ultrasound imaging analysis step;

[0134] The spatial variation of the frequency shift trend was calculated, the window size was set to 20×20 pixels, the difference in frequency shift stability within adjacent windows was calculated, the central difference method was used to calculate the frequency shift offset rate of the differentiated area, the frequency shift change trend of different areas was compared, the distribution characteristics of the offset change area were statistically analyzed, the offset change threshold was set, and 1.2 times the global offset rate mean was taken as the judgment benchmark. The offset rates of all pixel points were arranged, and the areas above the threshold were screened out. The data of all screened areas were integrated to finally obtain the ultrasound imaging analysis steps.

[0135] See also Figure 7 , a thyroid nodule ultrasound examination system, comprising:

[0136] The resonance mode extraction module obtains the ultrasonic echo signal of the thyroid tissue, analyzes the instantaneous amplitude, phase and short-time frequency characteristics of the echo signal in the nodule area, calculates the energy distribution and instantaneous frequency difference, and establishes the resonance mode feature matrix;

[0137] The signal enhancement control module calls the resonance mode characteristic matrix, calculates the modal energy density of the echo signal in the nodule area, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping according to the gain adjustment coefficient, counts the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor according to 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, calculates the grayscale gradient of the ultrasound image in the nodule area, constructs the gradient distribution matrix, calls the gradient change rate to calculate the transformation kernel matrix, adjusts the pixel contrast according to the transformation 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 imaging matrix, extracts the short-time frequency characteristics of the echo signal in the nodule area, calculates the instantaneous frequency peak, builds an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to screen abnormal areas, and obtains the echo frequency offset matrix;

[0140] The spatial trend analysis module calls the echo frequency offset matrix, counts the spatial distribution density of the abnormal frequency shift area 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.

[0141] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for ultrasonic examination of thyroid nodules, characterized in that: The following steps are involved: S1: Acquire the ultrasonic echo signal of thyroid tissue, analyze the instantaneous characteristics of the signal in the nodule area, calculate the energy distribution and instantaneous frequency difference, screen the stable resonance mode, and establish the resonance mode characteristic matrix; S2: calling the resonance modal characteristic matrix, calculating the modal energy density of the echo signal in the nodule area, calculating the gain adjustment coefficient based on the signal peak curve, performing nonlinear mapping according to the gain adjustment coefficient, calling the mutual correlation matrix to screen the correlation signal, performing dynamic weight suppression to adjust the signal strength, and obtaining an enhanced echo signal matrix; S3: calling the enhanced echo signal matrix, calculating the grayscale gradient of the nodule area, adjusting the pixel contrast, and obtaining a super-resolution enhanced image matrix; S4: calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the nodule area, calculating the frequency change, and obtaining the echo frequency offset matrix; S5: calling the echo frequency offset matrix, calculating the abnormal coverage ratio and the frequency shift stability, and obtaining the ultrasonic imaging analysis steps.

2. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The resonance modal feature matrix includes modal energy characteristics, resonance frequency distribution, and instantaneous amplitude parameters; the enhanced echo signal matrix includes signal gain adjustment coefficients, noise suppression weights, and target signal energy ratios; the super-resolution enhanced image matrix includes pixel contrast parameters, local gradient mapping values, and brightness balance 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 step includes abnormal area coverage, frequency offset stability, and nodule tissue frequency shift characteristics.

3. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The specific steps of obtaining the ultrasonic echo signal of thyroid tissue, analyzing the instantaneous characteristics of the signal in the nodule area, calculating the energy distribution and instantaneous frequency difference, screening the stable resonance mode, and establishing the resonance mode characteristic matrix are as follows: S101: Acquire thyroid tissue ultrasonic echo signals, analyze the instantaneous amplitude, phase and short-time frequency of the signals, calculate the amplitude change rate, phase offset and short-time frequency variation within the time window, and obtain an 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 band, calculate the instantaneous frequency of the frequency band according to the energy distribution, identify the instantaneous frequency difference of the differentiated frequency band, calculate the frequency change rate, analyze the frequency difference distribution between the frequency bands, determine the change trend of the frequency difference in the time dimension, and obtain the instantaneous frequency difference matrix; S103: calling the instantaneous frequency difference matrix, screening the resonance modes with stable frequency differences, extracting the signal modes that meet the threshold requirements, integrating the corresponding amplitude, phase and short-time frequency data, and establishing the resonance mode feature matrix.

4. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The specific steps of calling the resonance modal characteristic matrix, calculating the modal energy density of the echo signal in the nodule area, calculating the gain adjustment coefficient based on the signal peak curve, performing nonlinear mapping according to the gain adjustment coefficient, calling the mutual correlation matrix to screen the correlation signal, performing dynamic weight suppression to adjust the signal strength, and obtaining the enhanced echo signal matrix are as follows: S201: calling the resonance modal characteristic matrix, analyzing the modal energy density of the echo signal in the nodule area, extracting the energy value of the frequency band signal, calculating the energy mean of the unit area, and counting the energy fluctuation range, obtaining the modal energy peak value and the corresponding frequency, and obtaining 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 of the peak area 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 tissue modal signal, construct a cross-correlation matrix between signals, screen the signal that meets the correlation threshold, adjust the signal mapping parameters according to the correlation coefficient, screen the background signal that meets the constraint, and obtain the signal cross-correlation matrix; S203: calling the signal correlation matrix, calculating the signal correlation coefficient, calculating the dynamic suppression factor according to the correlation coefficient, adjusting the signal strength, performing dynamic weight suppression, integrating the adjusted signal matrix, and obtaining an enhanced echo signal matrix.

5. The method for ultrasonic examination of thyroid nodules according to claim 4, characterized in that: The energy mean calculation formula is specifically: Among them, E avg represents the mean modal energy per unit area, N represents the number of sampling points per unit area, and A i represents the real signal component of the i-th sampling point, B i represents the imaginary signal component of the i-th sampling point, C i represents the instantaneous frequency of the i-th sampling point, Represents the instantaneous frequency mean of all sampling points in the unit area.

6. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The specific steps of calling the enhanced echo signal matrix, calculating the gray gradient of the nodule area, adjusting the pixel contrast, and obtaining the super-resolution enhanced image matrix are as follows: S301: calling the enhanced echo signal matrix, calculating the grayscale gradient of the ultrasound image of the nodule area, extracting the grayscale change rate of each pixel in the image, calculating the gradient change amplitude between the pixels based on the gradient distribution, constructing the grayscale gradient distribution matrix, and obtaining the gradient distribution matrix; S302: Based on the gradient distribution matrix, analyze the gradient change rate, count the gradient change trend of the pixel points 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 points, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix; S303: calling the transformation kernel matrix, calculating the neighborhood brightness adjustment value in combination with the brightness gradient, extracting the neighborhood brightness change trend of the pixel point, identifying the brightness change coefficient, and correcting the image pixel brightness according to the adjustment coefficient, integrating the adjusted image data, and obtaining the super-resolution enhanced image matrix.

7. The method for ultrasonic examination of thyroid nodules according to claim 6, characterized in that: The gradient change trend of the pixels in the statistical image area is calculated using the formula: 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 point, adjust the image contrast distribution according to the change value, and obtain the transformation kernel matrix; Among them, ΔK represents the average gradient change rate percentage, M represents the total number of pixels in the calculation area, and G k Represents the gradient value of the kth pixel, G avg Represents the average value of all pixel gradient values ​​in the area, sum symbol Indicates the accumulation of gradient changes of all pixels.

8. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The specific steps of calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the nodule area, calculating the frequency change, and obtaining the echo frequency offset matrix are as follows: S401: calling the super-resolution enhanced image matrix, extracting the short-time frequency characteristics of the echo signal in the nodule area, calculating the instantaneous frequency value of the pixel point, obtaining the frequency change trend of the local area, identifying the instantaneous frequency peak, and generating the instantaneous frequency peak matrix; S402: Based on the instantaneous frequency peak matrix, the frequency change of the pixel point in the spatial position is calculated, the frequency change difference between the pixels is extracted, the change rate in the local frequency gradient direction is counted, the spatial frequency offset amplitude value is calculated, the local change trend of the offset distribution is analyzed, and the offset distribution matrix is ​​constructed according to the frequency change rate. The area where the offset is higher than the set threshold is screened to obtain the offset distribution matrix; S403: calling the offset distribution matrix, analyzing the variation characteristics of the spatial frequency gradient, calculating the degree of local frequency drift, screening the frequency offset amount in the abnormal area, integrating the frequency variation data, and obtaining the echo frequency offset matrix.

9. The method for ultrasonic examination of thyroid nodules according to claim 1, characterized in that: The specific steps of calling the echo frequency offset matrix, calculating the abnormal coverage ratio and the frequency shift stability, and obtaining the ultrasonic imaging analysis step are as follows: S501: calling the echo frequency shift matrix, counting the spatial distribution density of the abnormal frequency shift area inside the nodule, calculating the number ratio of the abnormal frequency shift points in the area, calculating the spatial coverage of the abnormal area according to the local distribution, and obtaining the abnormal frequency shift spatial density matrix; S502: Based on the abnormal frequency shift spatial density matrix, calculate the coverage ratio of the abnormal area within the nodule range, count the frequency shift value variation range of the coverage area, calculate the frequency fluctuation rate of the abnormal area, call the coverage ratio data to calculate the tissue frequency shift stability, calculate the stability coefficient according to the frequency shift mean of the differentiated area, screen the area with stability lower than the threshold, extract the abnormal frequency shift stability feature, calculate the frequency shift stability gradient, and obtain the tissue frequency shift stability matrix; S503: calling the tissue frequency shift stability matrix, calculating the spatial variation of the frequency shift trend, extracting the frequency shift offset rate of the differentiated area, statistically analyzing the distribution characteristics of the offset change area, integrating the regional frequency shift change data, and obtaining the ultrasound imaging analysis step.

10. A thyroid nodule ultrasound examination system, characterized in that: According to a method for ultrasonic examination of thyroid nodules according to any one of claims 1 to 9, the system comprises: The resonance mode extraction module obtains the ultrasonic echo signal of the thyroid tissue, analyzes the instantaneous amplitude, phase and short-time frequency characteristics of the echo signal in the nodule area, calculates the energy distribution and instantaneous frequency difference, and establishes the resonance mode feature matrix; The signal enhancement control module calls the resonance modal feature matrix, calculates the modal energy density of the echo signal in the nodule area, calculates the gain adjustment coefficient based on the signal peak curve, performs nonlinear mapping according to the gain adjustment coefficient, counts the cross-correlation value of the background tissue modal signal, calculates the dynamic suppression factor according to the signal cross-correlation coefficient, performs dynamic weight suppression to adjust the signal strength, and obtains an enhanced echo signal matrix; The image gradient optimization module calls the enhanced echo signal matrix, calculates the grayscale gradient of the ultrasound image of the nodule area, constructs a gradient distribution matrix, calls the gradient change rate to calculate the transformation kernel matrix, adjusts the pixel contrast according to the transformation 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 imaging matrix, extracts the short-time frequency characteristics of the echo signal in the nodule area, calculates the instantaneous frequency peak, constructs an offset distribution matrix based on the frequency offset amplitude, calls the offset distribution matrix to screen the abnormal area, and obtains the echo frequency offset matrix; The spatial trend analysis module calls the echo frequency offset matrix, counts the spatial distribution density of the abnormal frequency shift area 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.

Images