Radio frequency gold microneedle control system based on subcutaneous impedance identification

Abstract

The invention relates to the technical field of medical instruments, and provides a radio frequency gold microneedle control system based on subcutaneous impedance identification, and a radio frequency gold microneedle is provided with an electrode array for impedance tomography. Comprising the following steps: an impedance distribution reconstruction module generates reconstructed impedance distribution from a voltage signal of subcutaneous tissue collected by an electrode array through an impedance distribution reconstruction model based on a genetic algorithm; the impedance imaging module generates an impedance distribution image for reflecting subcutaneous tissue impedance distribution from the reconstructed impedance distribution through an impedance rate segmentation model based on a UNet architecture; the depth control module determines the recommended depth of the radio frequency gold microneedle according to the impedance distribution image; and the radio frequency energy control module adjusts control parameters of a control algorithm for controlling the radio frequency energy of the radio frequency gold microneedle according to the plurality of impedance distribution images. Accurate segmentation judgment and better imaging effect on impedance distribution are achieved, and then impedance imaging is applied to depth and energy pulse control of the radio frequency 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] Radio frequency gold microneedle, also known as gold microneedle radio frequency technology, combines microneedle and radio frequency technology. After the radio frequency gold microneedle penetrates into a specific layer of the skin epidermis, it emits radio frequency energy, allowing the radio frequency energy to reach the dermis below the epidermis, and then directly injects the energy into the dermis completely. When the tissue in the dermis is exposed to radio frequency energy, due to the impedance of the tissue to electromagnetic waves, the charged particles will vibrate and rub violently under the action of electromagnetic waves to generate heat, thereby stimulating the contraction and regeneration of collagen, promoting the effective denaturation, reorganization and proliferation of proteins, so as to achieve the rejuvenation and firming of facial skin, and improve crow's feet, forehead wrinkles, nasolabial folds, neck wrinkles, etc.

[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 voltage of the electrodes, 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 radio frequency gold microneedle treatment. The size of the impedance is affected by many factors, including the water content of the tissue, ion concentration, temperature, and the frequency of the radio frequency. In radio frequency gold microneedle treatment, the depth of microneedle insertion will directly affect the distribution of radio frequency energy in the subcutaneous tissue and the heating effect. As the insertion depth of the microneedle increases, the tissue layer that the radio frequency energy needs to pass through becomes thicker, and the electrical impedance also increases accordingly, resulting in a decrease in the transmission efficiency of radio frequency energy and a weakened heating effect. However, radio frequency energy can act more deeply on the subcutaneous tissue, stimulating the contraction and regeneration of deeper collagen, which is more helpful to improve skin sagging, wrinkles and other problems caused by deep tissues, and improve the treatment effect. Therefore, during the treatment of radio frequency gold microneedles, it is necessary to determine the appropriate depth and depth of radio frequency 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 radio frequency gold microneedle control system based on subcutaneous impedance recognition. Through 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 judgment of the impedance distribution and better imaging effect are achieved, thereby realizing the application of impedance imaging to the depth and energy pulse control of the radio frequency gold microneedle based on subcutaneous impedance recognition.

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

[0009] An impedance distribution reconstruction module, used for generating a reconstructed impedance distribution by using a genetic algorithm-based impedance distribution reconstruction model to transform the voltage signal of the subcutaneous tissue collected by the electrode array.

[0010] An impedance imaging module, used for generating an impedance distribution image reflecting the impedance distribution of subcutaneous tissue by using an impedance ratio segmentation model based on a UNet architecture to perform the reconstructed impedance distribution;

[0011] A depth control module, used to determine the recommended depth of the radio frequency 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] Further, 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 generation unit is used to generate an initial population of impedance distribution using an initialization strategy for the chromosome sequence;

[0016] The genetic operation unit is used to perform genetic operation on the initial population of impedance distribution 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 to 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 generation subunit is used 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 also 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 a better imaging effect.

[0024] Further, 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 used to segment the impedance characteristics of the extended impedance distribution through the impedance ratio segmentation model to generate the impedance distribution image.

[0027] Further, 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 calculate the improved hybrid loss function by performing weighted square sum on the cross entropy loss term and the Dice loss term.

[0031] In the above scheme, the impedance rate segmentation model based on the UNet architecture is improved to introduce the prior knowledge of the sample set to segment and stratify the impedance distribution. In particular, the improved hybrid loss function is combined with the global optimization ability of the cross entropy loss term and the local overlapping optimization ability of the Dice loss term, which amplifies the less obvious loss term differences in the impedance distribution of subcutaneous tissue, 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 used to determine a safe depth range that matches the set transmission efficiency of radio frequency energy according to the impedance distribution image;

[0034] 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.

[0035] Further, 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 used to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values ​​of a 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] Further, 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 used to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values ​​of a 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] Further, 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 more optimal impedance distribution imaging, and the dynamic PID control algorithm is used to achieve a fine, 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 beneficial effects:

[0046] 1. Through the global search capability of genetic algorithms and the introduction of prior knowledge of sample sets through radio frequency energy and UNet architecture, accurate segmentation judgment of impedance distribution and better imaging effects are achieved, thereby realizing the application of impedance imaging in the depth and energy pulse control of radio frequency gold microneedles based on subcutaneous impedance recognition.

[0047] 2. The voltage signals collected by the electrode array are 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 a better imaging effect.

[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 overlapping optimization capability of the Dice loss term, we have magnified 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 the safe depth range based on better impedance distribution imaging, and the dynamic PID control algorithm is used to achieve a more precise and sensitive and safer adjustment of the RF energy of the RF gold microneedle. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0051] Figure 2 It is a 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 It is a schematic diagram of the principle framework of impedance distribution image generation of the radio frequency gold microneedle control system based on subcutaneous impedance recognition according to an embodiment of the present invention;

[0053] Figure 4It is a structural schematic diagram of an impedance classification model based on the UNet architecture of 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 only used to explain the present invention and are not used 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 protection scope 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 drawings. This is merely 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] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to 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, which realizes accurate segmentation judgment of impedance distribution and better imaging effect through the global search capability of genetic algorithm and the prior knowledge of radio frequency energy and UNet architecture introduced into the sample set, thereby realizing the application of impedance imaging to the depth and energy pulse control of radio frequency gold microneedles based on subcutaneous impedance recognition.

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

[0060] A depth control module is used to determine the recommended depth of the RF gold microneedle according to the impedance distribution image; and a RF energy control module is used to adjust the control parameters of the control algorithm according to a plurality of the impedance distribution images at a set time interval, and 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 microneedles, and the array on the circle represents the electrode array. The electrode array is preferably arranged around the radio frequency gold microneedles. The electrode array is preferably arranged in a circle shape. An electromagnetic isolation cover is preferably arranged between the electrode array and the radio frequency gold microneedles 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 initial population of impedance distribution using an initialization strategy for the chromosome sequence; the genetic operation unit is used to perform genetic operations on the initial population of impedance distribution through the impedance distribution reconstruction model to generate the reconstructed impedance distribution.

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

[0064] Based on this, an effective initialization strategy is used for the chromosome sequence to generate a high-quality initial population of impedance distribution, specifically: the initial population of impedance distribution is generated by randomly arranging the operation order of the first layer impedance sequence and the second layer position sequence using a random number generator. Therefore, the diversity of the initial population can be broadened, and the generation of high-quality and diverse initial populations can ensure that the algorithm can quickly and effectively converge to the optimal solution, and then calculate a more realistic reconstructed impedance distribution through a genetic algorithm.

[0065] Furthermore, 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 to 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 generation subunit is used 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 equally into multiple impedance areas according to arcs of set angles, and 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] Furthermore, 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] In the formula, 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 direction similar to the target vector while maintaining a certain difference in numerical 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 in line with the situation that 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 a better imaging effect.

[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, Figure 4 As shown in the figure, 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, which are mapped to a new feature space through the nonlinear transformation capabilities of the hidden layer and the output layer to generate a new data representation and realize 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, Figure 4As shown, the impedance features are accurately segmented and graded through the one-hot encoding algorithm: the impedance level is determined as 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 determined by the vector representation is used to generate an impedance distribution image.

[0082] Therefore, the impedance distribution image realizes 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, so 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 weighting 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 ith 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 is improved to introduce the prior knowledge of the sample set to segment and stratify the impedance distribution. In particular, the improved hybrid loss function is combined with the global optimization ability of the cross entropy loss term and the local overlapping optimization ability of the Dice loss term, which amplifies the less obvious loss term differences in the impedance distribution of subcutaneous tissue, thereby achieving better imaging effects.

[0097] Furthermore, 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 the safe depth range that matches the set radio frequency energy transmission efficiency according to the impedance distribution image, which can be expressed as:

[0099] H max =H+Z max η,Z max η≤T 1

[0100] H min =HZ min η,Z min η≥T 2

[0101] In the formula, 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 is undergoing 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 are the maximum impedance value of the impedance distribution image and the minimum impedance value whose area is greater than the set limit value, T 1 、T 2 They are the upper safety value and the lower safety value, preferably 8*10-3 Ohm per square meter and 5*10 -5 Ohms per square meter.

[0102] It is understandable that during the radiofrequency treatment of radiofrequency gold microneedles, the impedance of the subcutaneous tissue may change as the treatment depth increases, and identifying the extreme value of impedance can help doctors understand the distribution of radiofrequency energy in the subcutaneous tissue. When radiofrequency energy encounters high-impedance tissue, the energy is concentrated in this tissue layer, causing the temperature of this layer of tissue to rise, 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] In the formula, H rec 、T rec Respectively represent the recommended depth and recommended working time, 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] Furthermore, 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 RF 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 RF energy of the RF gold microneedle through the dynamic PID control algorithm.

[0108] It can be understood that the RF gold microneedle control with a real-time intelligent feedback system (IFS) is constituted by the dynamic PID control algorithm. By real-time measurement of tissue impedance, the RF energy including RF frequency and RF power is feedback controlled to deliver constant energy to the tissue of the patient's subcutaneous dermis layer, maintain the temperature of the dermis layer tissue, make the treatment process independent of individual impedance changes, and achieve the purpose of accurately controlling tissue response. In addition, as the treatment progresses, the impedance characteristics of the subcutaneous tissue will change, and the dynamic PID parameter adjustment can automatically optimize the control parameters according to these changes, making the reaction process more stable and efficient.

[0109] Further, 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, and || means taking 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 plurality of impedance intermediate values, which can be expressed as:

[0114]

[0115] In the formula, K i ′, K′ d Respectively represent the adjusted differential coefficient and integral coefficient, K i , K d Respectively 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 more optimal impedance distribution imaging, and the dynamic PID control algorithm is used to achieve a fine, sensitivity-adjustable and safer adjustment of the RF energy of the RF gold microneedle.

[0117] In this embodiment, the global search capability of the genetic algorithm and the prior knowledge of the sample set introduced by the radio frequency energy and the UNet architecture achieve accurate segmentation judgment and better imaging effect for the impedance distribution, thereby achieving the application of impedance imaging to the depth and energy pulse control of the radio frequency gold microneedle based on subcutaneous impedance recognition. The genetic algorithm encodes the voltage signal collected by the electrode array in a set manner, generates the initial population, performs fitness evaluation and genetic operations, and achieves a more accurate impedance distribution through the global search capability of the genetic algorithm, thereby achieving a better imaging effect. The impedance rate segmentation model based on the UNet architecture is improved to achieve the segmentation and stratification of the impedance distribution by introducing the prior knowledge of the sample set, especially by improving the mixed loss function combined with the global optimization capability of the cross entropy loss term and the local overlap optimization capability of the Dice loss term, so that it amplifies the less obvious loss term difference in the impedance distribution of the subcutaneous tissue, thereby achieving a better imaging effect. The recommended depth of the radio frequency gold microneedle is calculated within the safe depth range according to the better impedance distribution imaging, and the radio frequency energy of the radio frequency gold microneedle is finely adjusted with adjustable sensitivity and safer through the dynamic PID control algorithm.

[0118] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0119] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope 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, used for generating a reconstructed impedance distribution by using a genetic algorithm-based impedance distribution reconstruction model to transform the voltage signal of the subcutaneous tissue collected by the electrode array. An impedance imaging module, used for generating an impedance distribution image reflecting the impedance distribution of subcutaneous tissue by using an impedance ratio segmentation model based on a UNet architecture to perform the reconstructed impedance distribution; A depth control module, used to determine the recommended depth of the radio frequency gold microneedle according to the impedance distribution image; 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.

2. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 1, characterized in that: 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 generation unit is used to generate an initial population of impedance distribution using an initialization strategy for the chromosome sequence; The genetic operation unit is used to perform genetic operation on the initial population of impedance distribution through the impedance distribution reconstruction model to generate the reconstructed impedance distribution.

3. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 2, characterized in that: 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 acquisition subunit is used to control each electrode in the electrode array to perform adjacent excitations, and to 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 sequence generation subunit is used 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; 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.

4. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 3, characterized in that: The impedance distribution reconstruction module also includes a fitness setting unit; 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.

5. 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 used to segment the impedance characteristics of the extended impedance distribution through the impedance ratio segmentation model to generate the impedance distribution image.

6. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 5, 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.

7. 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 calculate the improved hybrid loss function by performing weighted square sum on the cross entropy loss term and the Dice loss term.

8. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to any one of claims 1 to 7, characterized in that: The depth control module includes a depth range determination unit and a depth determination unit; The depth range determination unit is used to determine a safe depth range that matches the set transmission efficiency of radio frequency energy according to the impedance distribution image; The depth determination unit is used to calculate the recommended depth within the safe depth range according to the set working duration of the radio frequency energy.

9. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 8, 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 used to adjust the differential coefficient and the integral coefficient according to a plurality of impedance intermediate values ​​of a 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.

10. The radio frequency gold microneedle control system based on subcutaneous impedance recognition according to claim 9, 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 the impedance intermediate value according to 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