A radiofrequency gold microneedle control system based on subcutaneous impedance recognition

Abstract

The present invention relates to the field of medical device technology, and provides a radiofrequency gold microneedle control system based on subcutaneous impedance identification. The radiofrequency gold microneedle is equipped with an electrode array for impedance tomography imaging. The control system comprises: an impedance distribution reconstruction module that generates a reconstructed impedance distribution by applying a genetic algorithm-based impedance distribution reconstruction model to the voltage signal of the subcutaneous tissue collected by the electrode array; an impedance imaging module that generates an impedance distribution image reflecting the impedance distribution of the subcutaneous tissue by applying an impedance ratio segmentation model based on a UNet architecture to the reconstructed impedance distribution; a depth control module that determines the recommended depth of the radiofrequency gold microneedle based on the impedance distribution image; and a radiofrequency energy control module that adjusts the control parameters of the control algorithm for controlling the radiofrequency energy of the radiofrequency gold microneedle based on multiple impedance distribution images. The present invention achieves accurate segmentation and judgment of the impedance distribution and better imaging effects, thereby enabling the application of impedance imaging to control the depth and energy pulses of the radiofrequency gold microneedle.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and in particular to a radio frequency gold microneedle control system based on subcutaneous impedance recognition. Background Art

[0002] Radiofrequency gold microneedle therapy, also known as gold microneedle radiofrequency technology, combines microneedle and radiofrequency technologies. After penetrating deep into specific layers of the epidermis, the radiofrequency gold microneedle emits radiofrequency energy, allowing it to reach the dermis beneath the epidermis, directly injecting the energy into the dermis. When the dermis tissue is exposed to radiofrequency energy, due to the tissue's resistance to electromagnetic waves, the charged particles vibrate violently under the influence of electromagnetic waves, generating heat, which in turn stimulates the contraction and regeneration of collagen, promoting the effective denaturation, reorganization, and proliferation of proteins. This results in facial skin rejuvenation and firming, improving the appearance of crow's feet, forehead wrinkles, nasolabial folds, and neck lines.

[0003] Electrical impedance tomography (EIT) technology places multiple electrodes in a set shape on the facial surface, applies an appropriate and safe excitation current to the electrodes, collects the electrode voltage, and uses an image reconstruction algorithm to reconstruct an image of the impedance distribution inside biological tissues using the voltage and current data.

[0004] At present, the impedance of the subcutaneous tissue in the dermis is a key factor that needs to be considered in radiofrequency gold microneedle treatment. The magnitude of impedance is affected by many factors, including the water content of the tissue, ion concentration, temperature, and the frequency of radiofrequency. In radiofrequency gold microneedle treatment, the depth of microneedle insertion will directly affect the distribution of radiofrequency energy in the subcutaneous tissue and the heating effect. As the insertion depth of the microneedle increases, the tissue layer that the radiofrequency energy needs to pass through becomes thicker, and the electrical impedance also increases accordingly, resulting in a decrease in the transmission efficiency of radiofrequency energy and a weakening of the heating effect. However, radiofrequency energy can act more deeply on the subcutaneous tissue, stimulating deeper collagen contraction and regeneration, which is more conducive to improving skin sagging, wrinkles and other problems caused by deep tissues, and improving the treatment effect. Therefore, during the treatment of radiofrequency gold microneedles, it is necessary to determine the appropriate depth and depth of radiofrequency gold microneedles based on the impedance of the subcutaneous tissue.

[0005] Therefore, during the treatment process, doctors need to accurately adjust the insertion depth of the microneedles according to the patient's skin condition and needs to ensure that the radiofrequency energy can accurately act on the target tissue layer while avoiding damage to surrounding tissues.

[0006] Therefore, how to improve impedance imaging so that it can be applied to the depth and energy pulse control of radiofrequency gold microneedles based on subcutaneous impedance identification is a technical problem that needs to be solved. Summary of the Invention

[0007] To this end, the present invention provides a radiofrequency gold microneedle control system based on subcutaneous impedance recognition. By leveraging the global search capability of the genetic algorithm and the prior knowledge of radiofrequency energy and the UNet architecture introduced into the sample set, accurate segmentation and judgment of the impedance distribution and better imaging effects are achieved, thereby enabling impedance imaging to be applied to the depth and energy pulse control of radiofrequency gold microneedles based on subcutaneous impedance recognition.

[0008] To achieve the above objectives, the present invention proposes a radiofrequency gold microneedle control system based on subcutaneous impedance recognition, wherein the radiofrequency gold microneedle is provided with an electrode array for impedance tomography, comprising:

[0009] An impedance distribution reconstruction module, configured to generate a reconstructed impedance distribution by using a genetic algorithm-based impedance distribution reconstruction model to generate a reconstructed impedance distribution based on the voltage signal of the subcutaneous tissue collected by the electrode array;

[0010] An impedance imaging module, configured to generate an impedance distribution image reflecting the impedance distribution of subcutaneous tissue by applying the reconstructed impedance distribution to an impedance ratio segmentation model based on a UNet architecture;

[0011] a depth control module, configured to determine a recommended depth of the radiofrequency gold microneedle according to the impedance distribution image;

[0012] The radio frequency energy control module is used to adjust the control parameters of the control algorithm according to the plurality of impedance distribution images at set time intervals, and the control algorithm is used to control the radio frequency energy of the radio frequency gold microneedle.

[0013] Furthermore, the impedance distribution reconstruction module includes an encoding unit, an initial population generation unit and a genetic operation unit;

[0014] The encoding unit is used to perform chromosome encoding on the mapping impedances of the plurality of voltage signals according to their excitation position sequences to generate a chromosome sequence;

[0015] The initial population generating unit is used to generate an initial impedance distribution population using an initialization strategy for the chromosome sequence;

[0016] The genetic operation unit is used to perform genetic operation on the initial impedance distribution population through the impedance distribution reconstruction model to generate the reconstructed impedance distribution.

[0017] Furthermore, the voltage signal includes an excitation voltage signal, a ground voltage signal, and an average voltage signal; the chromosome sequence includes a first-layer voltage signal sequence and a second-layer position sequence; and the encoding unit includes a voltage acquisition encoding subunit, a voltage position encoding subunit, and a chromosome sequence generation subunit;

[0018] The voltage acquisition subunit is used to control each electrode in the electrode array to perform adjacent excitations and construct a voltage sequence corresponding to the excitation voltage signal, the ground voltage signal and the average voltage signal generated by all adjacent excitations;

[0019] The sequence generating subunit is configured to generate and construct a corresponding first-layer impedance sequence according to the type of the voltage signal, and to construct a corresponding second-layer position sequence according to the voltage sequence;

[0020] The chromosome sequence generating subunit is used to generate the chromosome sequence according to the first layer impedance sequence and the second layer position sequence.

[0021] Furthermore, the impedance distribution reconstruction module further includes a fitness setting unit;

[0022] The fitness setting unit is connected to the initial population generation unit and is used to construct a fitness function of the impedance distribution reconstruction model according to the difference between the cosine similarity and the Manhattan norm of the voltage sequence and the chromosome sequence.

[0023] In the above scheme, the voltage signal collected by the electrode array is encoded in a set manner through a genetic algorithm, an initial population is generated, fitness evaluation and genetic operations are performed, and a more accurate impedance distribution is generated through the global search capability of the genetic algorithm, thereby achieving better imaging effects.

[0024] Furthermore, the impedance imaging module includes a data expansion unit and a tomographic image generation unit;

[0025] The data expansion unit is used to perform data expansion on the reconstructed impedance distribution through a fully connected network to generate an extended impedance distribution;

[0026] The tomographic image generating unit is configured to segment the impedance characteristics of the extended impedance distribution using the impedance ratio segmentation model to generate the impedance distribution image.

[0027] Furthermore, the tomographic image generation unit includes a judgment subunit;

[0028] The judgment subunit is used to segment the impedance features of the output layer of the impedance ratio segmentation model through a one-hot encoding algorithm to generate the impedance distribution image.

[0029] Furthermore, the impedance ratio segmentation model adopts an improved hybrid loss function;

[0030] The improved hybrid loss function is used to perform weighted square sum calculation on the cross entropy loss term and the Dice loss term to calculate the improved hybrid loss function.

[0031] In the above scheme, the impedance rate segmentation model based on the UNet architecture was improved to introduce the prior knowledge of the sample set to segment and stratify the impedance distribution. In particular, by improving the hybrid loss function and combining the global optimization capability of the cross-entropy loss term and the local overlapping optimization capability of the Dice loss term, the less obvious loss term differences in the impedance distribution of subcutaneous tissue were amplified, thereby achieving better imaging effects.

[0032] Further, the depth control module includes a depth range determination module and a depth determination module;

[0033] The depth range determination module is configured to determine a safe depth range that matches a set transmission efficiency of radio frequency energy based on the impedance distribution image;

[0034] The depth determination module is configured to calculate the recommended depth within the safe depth range according to the set working duration of the radio frequency energy.

[0035] Furthermore, the control algorithm is a dynamic PID control algorithm, the control parameters are differential coefficients and integral coefficients, and the radio frequency energy control module includes an adjustment unit and a control unit;

[0036] The adjustment unit is configured to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values of the plurality of impedance distribution images;

[0037] The control unit is used to control the radio frequency energy of the radio frequency gold microneedle through the dynamic PID control algorithm.

[0038] Furthermore, the control algorithm is a dynamic PID control algorithm, the control parameters are differential coefficients and integral coefficients, and the radio frequency energy control module includes an adjustment unit and a control unit;

[0039] The adjustment unit is configured to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values of the plurality of impedance distribution images;

[0040] The control unit is used to control the radio frequency energy of the radio frequency gold microneedle through the dynamic PID control algorithm.

[0041] Furthermore, the adjustment unit includes an intermediate value calculation subunit and an adjustment amount calculation subunit;

[0042] The intermediate value calculation subunit is used to calculate the impedance intermediate value according to the impedance median and median area impedance reflected by the impedance distribution image;

[0043] The adjustment amount calculation subunit is used to determine the adjustment amounts of the differential coefficient and the integral coefficient according to the difference between the plurality of impedance intermediate values.

[0044] In the above scheme, the recommended depth of the RF gold microneedle is calculated within the safe depth range based on the better impedance distribution imaging, and the dynamic PID control algorithm is used to achieve a finer, sensitivity-adjustable and safer adjustment of the RF energy of the RF gold microneedle.

[0045] Compared with the prior art, the present invention has the following advantages:

[0046] 1. By leveraging the global search capabilities of genetic algorithms, radio frequency energy, and the UNet architecture to introduce prior knowledge of sample sets, we achieved accurate segmentation and better imaging of impedance distribution, thereby enabling the application of impedance imaging to control the depth and energy pulses of radio frequency gold microneedles based on subcutaneous impedance recognition.

[0047] 2. The genetic algorithm is used to encode the voltage signals collected by the electrode array in a set manner, generate the initial population, perform fitness evaluation and genetic operations, and achieve a more accurate impedance distribution through the global search capability of the genetic algorithm, thereby achieving better imaging effects.

[0048] 3. By improving the impedance segmentation model based on the UNet architecture, we have achieved the segmentation and stratification of the impedance distribution by introducing the prior knowledge of the sample set. In particular, by improving the hybrid loss function and combining the global optimization capability of the cross-entropy loss term and the local overlap optimization capability of the Dice loss term, we can amplify the less obvious loss term differences in the impedance distribution of subcutaneous tissue, thereby achieving better imaging effects.

[0049] 4. The recommended depth of the RF gold microneedle is calculated within a safe depth range based on better impedance distribution imaging, and a dynamic PID control algorithm is used to achieve a more precise and sensitive adjustment of the RF energy of the RF gold microneedle. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the structure of a radio frequency gold microneedle control system based on subcutaneous impedance recognition according to an embodiment of the present invention;

[0051] Figure 2 Schematic diagram of the impedance distribution image generation process of the radio frequency gold microneedle control system based on subcutaneous impedance recognition according to an embodiment of the present invention;

[0052] Figure 3 Schematic diagram of the impedance distribution image generation principle framework of the radio frequency gold microneedle control system based on subcutaneous impedance recognition according to an embodiment of the present invention;

[0053] Figure 4This is a structural diagram of an impedance classification model based on a UNet architecture for a radio frequency gold microneedle control system based on subcutaneous impedance recognition according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0055] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0056] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0057] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0058] like Figures 1 to 4 As shown, the present invention provides a radio frequency gold microneedle control system based on subcutaneous impedance recognition. By using the global search capability of the genetic algorithm and the prior knowledge of the sample set introduced by the radio frequency energy and UNet architecture, accurate segmentation and judgment of the impedance distribution and better imaging effects are achieved, thereby realizing the application of impedance imaging in the depth and energy pulse control of the radio frequency gold microneedle based on subcutaneous impedance recognition.

[0059] like Figures 1 to 4As shown, this embodiment proposes a radiofrequency gold microneedle control system based on subcutaneous impedance recognition. The radiofrequency gold microneedle is provided with an electrode array for impedance tomography imaging, including: an impedance distribution reconstruction module for generating a reconstructed impedance distribution by applying a voltage signal of subcutaneous tissue collected by the electrode array to an impedance distribution reconstruction model based on a genetic algorithm; an impedance imaging module for generating an impedance distribution image reflecting the impedance distribution of subcutaneous tissue by applying an impedance ratio segmentation model based on a UNet architecture to the reconstructed impedance distribution;

[0060] A depth control module is used to determine the recommended depth of the RF gold microneedle based on the impedance distribution image; and a RF energy control module is used to adjust the control parameters of a control algorithm based on a plurality of the impedance distribution images at set time intervals, wherein the control algorithm is used to control the RF energy of the RF gold microneedle.

[0061] It should be noted that, see Figure 3 In the figure, the array in the circle represents the radio frequency gold microneedle, and the array on the circle represents the electrode array. The electrode array is preferably arranged around the radio frequency gold microneedle, and the electrode array is preferably arranged in a circular shape. An electromagnetic isolation cover is preferably set between the electrode array and the radio frequency gold microneedle to avoid electromagnetic interference from each other.

[0062] Furthermore, if Figure 2 As shown, the impedance distribution reconstruction module includes an encoding unit, an initial population generation unit and a genetic operation unit; the encoding unit is used to perform chromosome encoding on the mapped impedances of the multiple voltage signals according to their excitation position sequence to generate a chromosome sequence; the initial population generation unit is used to generate an impedance distribution initial population using an initialization strategy for the chromosome sequence; the genetic operation unit is used to perform genetic operations on the impedance distribution initial population through the impedance distribution reconstruction model to generate the reconstructed impedance distribution.

[0063] It is understandable that the electrode array field of electrical impedance tomography (EIT) satisfies Maxwell's equations and electromagnetic field theory. Therefore, image reconstruction based on impedance tomography can be regarded as an exploration of the mapping relationship between the relative conductivity (i.e., impedance) distribution within the field and the voltage signal distribution.

[0064] Based on this, an effective initialization strategy is employed for the chromosome sequence to generate a high-quality initial population for the impedance distribution. Specifically, this strategy involves using a random number generator to randomly permutate the order of operations in the first-layer impedance sequence and the second-layer position sequence. This strategy broadens the diversity of the initial population. Generating a high-quality, diverse initial population ensures that the algorithm converges quickly and efficiently to the optimal solution, thereby enabling the genetic algorithm to calculate a more realistic reconstructed impedance distribution.

[0065] Further, if Figure 2 As shown, the voltage signal includes an excitation voltage signal, a ground voltage signal and an average voltage signal, the chromosome sequence includes a first-layer voltage signal sequence and a second-layer position sequence, and the encoding unit includes a voltage acquisition encoding subunit, a voltage position encoding subunit and a chromosome sequence generation subunit;

[0066] The voltage acquisition subunit is used to control each electrode in the electrode array to perform adjacent excitations and construct a voltage sequence corresponding to the excitation voltage signal, the ground voltage signal and the average voltage signal generated by all adjacent excitations;

[0067] The sequence generating subunit is configured to generate and construct a corresponding first-layer impedance sequence according to the type of the voltage signal, and to construct a corresponding second-layer position sequence according to the voltage sequence;

[0068] The chromosome sequence generating subunit is used to generate the chromosome sequence according to the first layer impedance sequence and the second layer position sequence.

[0069] Specifically, the process of adjacent excitation is as follows: first, clarify the excitation and measurement methods: for example, there are 10 electrodes, marked as 1 to 10; after obtaining the ground voltage signal including all electrodes, select two adjacent electrodes as the excitation electrode pair each time, obtain 1 excitation voltage signal of the excitation electrode pair and the average voltage signal of 8 other electrodes, and then make the 10 electrodes serve as excitation electrodes in turn. Therefore, the electrode at one position can obtain 1 ground voltage signal, 1 excitation voltage signal and 8 average voltage signals, a total of 10 voltage signals, and a total voltage sequence of 100 voltage signals is obtained.

[0070] Specifically, the first layer impedance sequence and the second layer position sequence are located in two layers of vector models respectively. The first layer impedance sequence is as follows: Figure 3 The circled area shown is divided into multiple impedance areas in equal parts according to arcs of set angles. The second-layer position sequence is the order of impedance areas that is the same as the adjacent excitation order represented by the voltage sequence, so that it can generate a chromosome sequence with multiple possible solutions that conforms to the format through encoding.

[0071] Further, if Figure 2 As shown, the fitness setting unit is connected to the initial population generation unit to construct the fitness function of the impedance distribution reconstruction model according to the difference between the cosine similarity and the Manhattan norm of the voltage sequence and the chromosome sequence.

[0072] Specifically, the fitness function is expressed as:

[0073]

[0074] Where f represents the fitness function, P represents the chromosome sequence, V represents the voltage sequence, || represents the cosine similarity of the sequence, which is mapped to the non-negative interval, and || || represents the Manhattan norm. Squaring it can amplify the difference while preventing it from falling into the negative interval.

[0075] Therefore, the fitness function combines cosine similarity and Manhattan norm, where cosine similarity focuses on the direction of the vector and Manhattan norm measures the sum of the absolute differences of the vector, so that the genetic algorithm can find a solution with a similar direction to the target vector while maintaining a certain difference in value from the target vector. By linearly combining cosine similarity and Manhattan norm, a slightly fault-tolerant reconstructed impedance distribution can be achieved, which is consistent with the situation where subcutaneous tissue impedance is affected by multiple factors.

[0076] Specifically, the Pareto-based genetic algorithm framework.

[0077] In the above scheme, the voltage signal collected by the electrode array is encoded in a set manner through a genetic algorithm, an initial population is generated, fitness evaluation and genetic operations are performed, and a more accurate impedance distribution is generated through the global search capability of the genetic algorithm, thereby achieving better imaging effects.

[0078] Furthermore, if Figure 4 As shown, the impedance imaging module includes a data expansion unit and a tomographic image generation unit; the data expansion unit is used to perform data expansion on the reconstructed impedance distribution through a fully connected network to generate an extended impedance distribution; the tomographic image generation unit is used to perform impedance feature segmentation on the extended impedance distribution through the impedance ratio segmentation model to generate the impedance distribution image.

[0079] Specifically, if Figure 4 As shown in Figure 1, the number of nodes in the input layer, hidden layer, and output layer of the fully connected network (FCN) are 5, 6, and 8, respectively. The input layer is used as a feature extractor to extract useful features from the input reconstructed impedance distribution. The nonlinear transformation capabilities of the hidden layer and output layer are used to map them to a new feature space to generate a new data representation, thereby achieving data expansion.

[0080] Furthermore, if Figure 4 As shown, the tomographic image generation unit includes a judgment subunit; the judgment subunit is used to segment the impedance characteristics of the output layer of the impedance ratio segmentation model through a one-hot encoding algorithm to generate the impedance distribution image.

[0081] Specifically, if Figure 4As shown, the impedance features are accurately segmented and graded through the one-hot encoding algorithm: the impedance level is determined to be one of very low, low, medium, high, and very high, and a vector representation is performed through the one-hot encoding algorithm, and the pixel value of the vector representation is determined to generate an impedance distribution image.

[0082] Therefore, the impedance distribution image enables the visualization of impedance data.

[0083] Furthermore, the impedance segmentation model adopts an improved hybrid loss function; the improved hybrid loss function is used to calculate the improved hybrid loss function by performing weighted square sum on the cross entropy loss term and the Dice loss term.

[0084] It is understandable that the improved hybrid loss function can also be replaced by a conventional loss function, which is more direct and insensitive to the difference between the predicted distribution and the true distribution, and thus the optimization effect is slightly worse, which is expressed as:

[0085] L C =α·L CE +β·L Dice

[0086] Where, L C Represents the improved hybrid loss function, α and β represent two weighted coefficients, L CE represents the cross entropy loss term, L Dice Represents the Dice loss term.

[0087] Specifically, the improved hybrid loss function is expressed as:

[0088]

[0089] Where, L C represents the improved hybrid loss function, α and β represent two weighting coefficients, preferably 0.6 and 0.4 to make the difference between the predicted distribution and the true distribution more prominent, L CE represents the cross entropy loss term, L Dice Represents the Dice loss term.

[0090] The expression of the cross entropy loss term is:

[0091]

[0092] Where, L CE represents the cross entropy loss term, N represents the total number of pixels, C represents the number of categories, and y i,c represents the true label of the cth category of the i-th pixel, represents the predicted label of the cth category of the i-th pixel. Therefore, the cross entropy loss term measures the classification difference between the predicted distribution and the true distribution.

[0093] The expression of Dice loss term is:

[0094]

[0095] Where, L Dice represents the Dice loss term, N represents the total number of pixels, y i represents the true label of the i-th pixel, represents the predicted label of the i-th pixel, represents the predicted label of the im+1th pixel, and ε is a smoothing term to prevent the denominator from taking zero. Therefore, the Dice loss term is used to measure the overlap between the predicted distribution and the true distribution. The closer the Dice loss term is to 1, the closer the predicted result is to the true value.

[0096] In the above scheme, the impedance rate segmentation model based on the UNet architecture was improved to introduce the prior knowledge of the sample set to segment and stratify the impedance distribution. In particular, by improving the hybrid loss function and combining the global optimization capability of the cross-entropy loss term and the local overlapping optimization capability of the Dice loss term, the less obvious loss term differences in the impedance distribution of subcutaneous tissue were amplified, thereby achieving better imaging effects.

[0097] Further, if Figure 3 As shown, the depth control module includes a depth range determination module and a depth determination module;

[0098] The depth range determination module is used to determine a safe depth range that matches the set radio frequency energy transmission efficiency based on the impedance distribution image, which can be expressed as:

[0099] H max =H+Z max η,Z max η≤T1

[0100] H min =HZ min η,Z min η≥T2

[0101] Where H min 、H max are the lower and upper limits of the safety depth range, respectively. H is the set depth, which is determined by the doctor based on whether the patient undergoes superficial treatment, middle treatment, or deep treatment, and the treatment method, and is between 0.5 mm and 5.0 mm. Z max , Z min The impedance distribution image has the maximum impedance value and the impedance minimum value whose area is greater than the set limit respectively. T1 and T2 are the upper limit safety value and the lower limit safety value respectively, preferably 8*10 -3 Ohms per square meter and 5*10 -5 Ohms per square meter.

[0102] It is understandable that during RF gold microneedle treatment, the impedance of the subcutaneous tissue may change as the treatment depth increases. Identifying the extreme impedance values can help doctors understand the distribution of RF energy in the subcutaneous tissue. This can prevent the RF energy from concentrating in high-impedance tissue when it encounters it, leading to increased temperature, overheating, and potential damage.

[0103] The depth determination module is used to calculate the recommended depth within the safe depth range according to the set working duration of the radio frequency energy, which can be expressed as:

[0104] H rec =H min +(H max -H min )(T w -T rec )

[0105] Where H rec 、T rec Represent the recommended depth and recommended working time respectively, H min 、H max are the lower and upper limits of the safety depth range, T w Indicates the duration of the work.

[0106] Therefore, the above scheme can finely adjust and control the depth of the radiofrequency gold microneedle set by the doctor according to the impedance distribution image.

[0107] Further, if Figure 3 As shown, the control algorithm is a dynamic PID control algorithm, the control parameters are differential coefficients and integral coefficients, and the radio frequency energy control module includes an adjustment unit and a control unit; the adjustment unit is used to adjust the differential coefficient and the integral coefficient according to multiple impedance intermediate values of the multiple impedance distribution images; the control unit is used to control the radio frequency energy of the radio frequency gold microneedle through the dynamic PID control algorithm.

[0108] It can be understood that the dynamic PID control algorithm constitutes a real-time intelligent feedback system (IFS) for radiofrequency gold microneedle control. By measuring tissue impedance in real time, feedback control of radiofrequency energy, including radiofrequency frequency and radiofrequency power, is used to deliver constant energy to the patient's subcutaneous dermal tissue, maintaining the temperature of the dermal tissue. This makes the treatment process independent of individual impedance changes, achieving the purpose of precisely controlling tissue response. In addition, as the treatment progresses, the impedance characteristics of the subcutaneous tissue will change. Dynamic PID parameter adjustment can automatically optimize the control parameters based on these changes, making the response process more stable and efficient.

[0109] Furthermore, the adjustment unit includes an intermediate value calculation subunit and an adjustment amount calculation subunit;

[0110] The intermediate value calculation subunit is used to calculate the impedance intermediate value according to the impedance median and median area impedance reflected by the impedance distribution image, which can be expressed as:

[0111]

[0112] Where p′ med 、p med 、p meds They represent the median impedance value, median impedance ratio, and median area impedance, respectively. || indicates the absolute value.

[0113] The adjustment amount calculation subunit is used to determine the adjustment amount of the differential coefficient and the integral coefficient according to the difference between the multiple impedance intermediate values, which can be expressed as:

[0114]

[0115] Where K i ′, K′ d Represent the adjusted differential coefficient and integral coefficient, K i , K d Represent the adjusted differential coefficient and integral coefficient, ΔK i , ΔK d Respectively represent the adjustment amount of the differential coefficient and the adjustment amount of the integral coefficient, p' medi -p′ medi-1 represents the difference between two intermediate impedance values, R is the number of intermediate impedance values, 0.1 and 0.125 are adjustment coefficients to make the dynamic PID more sensitive, and the differential coefficient and the integral coefficient are adjusted proportionally.

[0116] In the above scheme, the recommended depth of the RF gold microneedle is calculated within the safe depth range based on the better impedance distribution imaging, and the dynamic PID control algorithm is used to achieve a finer, sensitivity-adjustable and safer adjustment of the RF energy of the RF gold microneedle.

[0117] In this embodiment, the global search capabilities of a genetic algorithm, combined with prior knowledge of the sample set using radiofrequency energy and a UNet architecture, enable accurate segmentation and determination of impedance distribution and improved imaging results. This allows impedance imaging to be applied to control the depth and energy pulses of radiofrequency gold microneedles based on subcutaneous impedance recognition. By encoding the voltage signals collected by the electrode array in a predefined manner, generating an initial population, performing fitness evaluation, and genetic manipulation, the genetic algorithm's global search capabilities enable the generation of more accurate impedance distributions, thereby achieving better imaging results. By improving the impedance ratio segmentation model based on the UNet architecture, prior knowledge of the sample set is incorporated to segment and stratify the impedance distribution. In particular, an improved hybrid loss function, combining the global optimization capabilities of the cross-entropy loss term and the local overlap optimization capabilities of the Dice loss term, amplifies relatively subtle loss term differences in the impedance distribution of subcutaneous tissue, thereby achieving better imaging results. Based on the optimized impedance distribution imaging, a recommended depth for radiofrequency gold microneedles is calculated within a safe depth range. Furthermore, a dynamic PID control algorithm enables precise, sensitive, and safer adjustment of the radiofrequency energy of the radiofrequency gold microneedles.

[0118] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0119] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A radio frequency gold microneedle control system based on subcutaneous impedance recognition, characterized in that: The radiofrequency gold microneedle is equipped with an electrode array for impedance tomography, including: An impedance distribution reconstruction module, configured to generate a reconstructed impedance distribution by using a genetic algorithm-based impedance distribution reconstruction model to generate a reconstructed impedance distribution based on the voltage signal of the subcutaneous tissue collected by the electrode array; An impedance imaging module, configured to generate an impedance distribution image reflecting the impedance distribution of subcutaneous tissue by applying the reconstructed impedance distribution to an impedance ratio segmentation model based on a UNet architecture; a depth control module, configured to determine a recommended depth of the radiofrequency gold microneedle according to the impedance distribution image; A radio frequency energy control module, configured to adjust control parameters of a control algorithm according to a plurality of impedance distribution images at set time intervals, wherein the control algorithm is used to control the radio frequency energy of the radio frequency gold microneedles; The impedance distribution reconstruction module includes an encoding unit, an initial population generation unit and a genetic operation unit; The encoding unit is used to perform chromosome encoding on the mapping impedances of the plurality of voltage signals according to their excitation position sequences to generate a chromosome sequence; The initial population generating unit is used to generate an initial impedance distribution population using an initialization strategy for the chromosome sequence; The genetic operation unit is used to perform genetic operation on the initial impedance distribution population to generate the reconstructed impedance distribution; The voltage signal includes an excitation voltage signal, a ground voltage signal and an average voltage signal, the chromosome sequence includes a first-layer voltage signal sequence and a second-layer position sequence, and the encoding unit includes a voltage acquisition encoding subunit, a voltage position encoding subunit and a chromosome sequence generation subunit; The voltage position encoding subunit is used to control each electrode in the electrode array to perform adjacent excitations and construct a voltage sequence corresponding to the excitation voltage signal, the ground voltage signal and the average voltage signal generated by all adjacent excitations; The chromosome sequence generating subunit is configured to generate and construct a corresponding first-layer impedance sequence according to the type of the voltage signal, and to construct a corresponding second-layer position sequence according to the voltage sequence, and to generate the chromosome sequence according to the first-layer impedance sequence and the second-layer position sequence; The impedance distribution reconstruction module also includes a fitness setting unit; The fitness setting unit is used to construct a fitness function of the impedance distribution reconstruction model according to the difference between the cosine similarity and the Manhattan norm of the voltage sequence and the chromosome sequence.

2. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 1, characterized in that: The impedance imaging module includes a data expansion unit and a tomographic image generation unit; The data expansion unit is used to perform data expansion on the reconstructed impedance distribution through a fully connected network to generate an extended impedance distribution; The tomographic image generating unit is configured to segment the impedance characteristics of the extended impedance distribution using the impedance ratio segmentation model to generate the impedance distribution image.

3. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 2, characterized in that: The tomographic image generation unit includes a judgment subunit; The judgment subunit is used to segment the impedance features of the output layer of the impedance ratio segmentation model through a one-hot encoding algorithm to generate the impedance distribution image.

4. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 1, characterized in that: The impedance ratio segmentation model adopts an improved hybrid loss function; The improved hybrid loss function is used to perform weighted square sum calculation on the cross entropy loss term and the Dice loss term to determine the improved hybrid loss function.

5. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to any one of claims 1 to 4, characterized in that: The depth control module includes a depth range determination module and a depth determination module; The depth range determination module is configured to determine a safe depth range that matches a set transmission efficiency of radio frequency energy based on the impedance distribution image; The depth determination module is configured to calculate and determine the recommended depth within the safe depth range according to the set working duration of the radio frequency energy.

6. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 5, characterized in that: The control algorithm is a dynamic PID control algorithm, the control parameters are differential coefficients and integral coefficients, and the radio frequency energy control module includes an adjustment unit and a control unit; The adjustment unit is configured to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values of the plurality of impedance distribution images; The control unit is used to control the radio frequency energy of the radio frequency gold microneedle through the dynamic PID control algorithm.

7. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 6, characterized in that: The adjustment unit includes an intermediate value calculation subunit and an adjustment amount calculation subunit; The intermediate value calculation subunit is used to calculate and determine the impedance intermediate value based on the impedance median and median area impedance reflected by the impedance distribution image; The adjustment amount calculation subunit is used to determine the adjustment amounts of the differential coefficient and the integral coefficient according to the difference between the plurality of impedance intermediate values.

Images