Wearable breast cancer intelligent screening system

Abstract

The invention relates to a wearable breast cancer intelligent screening system which comprises a wearable detection unit which is symmetrically designed on two sides, senses resistance change under the action of external force or pressure through an integrated flexible sensing film, collects voltage through a sampling chip, calculates resistance and sends the resistance to a main control processing unit; the main control processing unit obtains the resistance data, controls transmitting and receiving antenna pairs in the antenna array and measures S21 parameters of the transmitting and receiving antenna pairs; and the PC generates a mammary gland internal dielectric characteristic image without deformation errors through various algorithms, and screens the breast cancer according to the image. According to the method, the flexible sensing technology and the microwave imaging technology are deeply fused, so that real-time sensing and self-adaptive correction of the geometric attitude of the antenna array are realized, and the accuracy and repeatability of a detection result are ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of breast cancer screening, and in particular to a wearable breast cancer intelligent screening system. BACKGROUND

[0002] Breast cancer is one of the most common malignant tumors in women worldwide, and early diagnosis is the key to improving the cure rate and reducing mortality. The current commonly used screening methods include mammography (molybdenum target), ultrasonic examination, MRI and clinical palpation, etc. Although the traditional imaging method has high imaging accuracy, it has certain limitations, such as: X-ray photography: requires special equipment and radiation protection facilities, and the diagnostic sensitivity for dense breast tissue is low; MRI: high imaging accuracy, but expensive equipment and long examination time; ultrasound: convenient and no radiation, but the results depend on the experience of the operator and the repeatability is poor.

[0003] In recent years, microwave imaging (Microwave Imaging, MWI) has become a research hotspot for early screening of breast cancer due to its advantages of no radiation, wearability, low cost, and repeatable detection. This method uses the difference in dielectric properties between breast tumors and normal tissues to collect microwave scattering signals through a multi-antenna array and perform inversion imaging. However, existing microwave imaging screening systems are mostly large fixed devices, which still have great deficiencies in portability and comfort. In order to improve clinical usability, research has gradually developed from fixed large platforms to wearable and flexible directions, that is, through a bra or patch structure, the antenna array is directly arranged on the surface of the breast to realize rapid deployment and detection. However, the flexible wearable device improves the convenience and comfort of the device, but also introduces errors caused by array displacement and deformation.

[0004] Currently, the prior art mainly falls into two categories: 1. Flexible wearable breast detection device: Typical such as wearable bra type equipment based on ultrasonic wave or infrared thermal imaging. This kind of device uses flexible sensing elements close to the surface of the breast for detection, has a certain portability and comfort, but due to the limitation of detection principle, its spatial resolution and tissue discrimination ability are insufficient, which is difficult to meet the accuracy requirements of early breast cancer screening. 2. Fixed microwave imaging system: usually composed of fixed microwave modules and external transmitting sources (such as horn antennas), this kind of system uses the difference in dielectric properties between breast tumors and normal tissues for imaging, and the imaging mechanism is suitable for breast cancer detection. But this kind of device is bulky and depends on fixed scene, which cannot meet the needs of wearability and daily rapid screening. Therefore, the prior art either has wearability but lacks imaging accuracy (ultrasound, thermal imaging bra), or has imaging accuracy but lacks portability (fixed microwave imaging), none of which can simultaneously consider flexible fitting detection and high-precision microwave imaging, and the antenna array inevitably bends, compresses or shifts during wearing, geometric errors cause signal deviation, which reduces the imaging quality, and the prior art cannot collect the shape of the antenna array in real time, so it cannot perform geometric correction. SUMMARY

[0005] The purpose of the present application is to overcome the shortcomings of the prior art and provide a wearable breast cancer intelligent screening system to solve the problems existing in the prior art.

[0006] The purpose of the present application is achieved by the following technical solutions: A wearable breast cancer intelligent screening system, which comprises a double-sided symmetrical wearable detection unit and a main control processing unit connected thereto, and the main control processing unit is connected with a PC.

[0007] The wearable detection unit is configured to sense the resistance change when the integrated flexible sensing film is subjected to external force or pressure, and to send the resistance calculated by the sampling chip to the main control processing unit;

[0008] The main control processing unit is configured to perform the following functions: a) obtaining the resistance data sent by the wearable detection unit, which represents the geometric pose thereof; b) controlling the antenna path switching module to select the transmitting antenna and receiving antenna pair in the antenna array, and measuring the S21 parameter between the antenna pair by the network analysis module; c) by controlling the switch sequence, realizing multi-path polling scanning of different sensor unit resistance data and different antenna pair S21 parameters;

[0009] The PC is configured to convert the acquired data into a curvature field matrix describing the bending degree of the antenna array through a resistance-curvature mapping algorithm, obtain geometric posture data describing the three-dimensional spatial form of the antenna array at the wearing moment through a surface reconstruction algorithm, correct the antenna coordinate parameters in the microwave imaging inversion algorithm through the geometric posture data, generate a breast internal dielectric property image eliminating deformation errors, and screen breast cancer according to the image.

[0010] The wearable detection unit includes two detection components symmetrically arranged on the left and right sides, respectively, for adhering to the bilateral breasts, and each side of the detection component includes a flexible antenna, an integrated flexible sensing film, two array gating chips, and a sampling chip.

[0011] The integrated flexible sensing film includes a plurality of array units arranged in an array, which are arranged at the positions of the bra adhering to the breasts, and the flexible antennas are arranged in an array between each row and each column of the array units, each row of the flexible antennas is connected with each row of the array units, and each column of the flexible antennas is connected with each column of the array units.

[0012] The two array gating chips are connected with each row of the flexible antennas and each column of the flexible antennas, respectively, and the two array gating chips are connected with the sampling chip.

[0013] The integrated flexible sensing film includes a protective layer made of polydimethylsiloxane material on the upper and lower sides and an AGNW layer in the middle; the AGNW layer is deposited and annealed into a film, and then precisely patterned into an n×n square array unit by laser ablation;

[0014] Each array unit is a resistance sensor, and when subjected to external force bending or pressure, the contact resistance of the AGNW layer changes reversibly; when the film is bent or stretched, the contact points between the silver nanowires decrease, and the nanowire network structure is rearranged, resulting in an increase in the contact resistance; when the film is compressed, the nanowire contact is more closely, the contact area increases, and the contact resistance decreases.

[0015] The two array gating chips realize fast switching of multi-channel analog signals, and each array unit collects voltage to calculate resistance through the sampling chip and sends it to the main control processing unit.

[0016] The main control processing unit includes an antenna path switching module, a network analysis module, an MCU, and a power module; the wearable detection unit is connected with the antenna path switching module, the switching sequence of the antenna path switching module is controlled by the MCU, the dynamic switching between the wearable detection unit and the main control processing unit is realized, and then multi-path polling scanning is realized; the network analysis module is connected with the antenna path switching module and the MCU, and is used for measuring the S21 parameters transmitted under different channel combinations; the power module provides power supply for the MCU, the antenna path switching module, and the network analysis module.

[0017] The resistance-curvature mapping algorithm comprises a resistance-curvature real-time mapping step, which comprises the following:

[0018] At a certain channel signal acquisition moment, the resistance value t_k, the posture data matrix R_k collected and packaged by the MCU, and each element R_k(i,j) in the matrix representing the instantaneous resistance value of the (i,j) sensing unit;

[0019] The posture data matrix R_k is input into the mapping function f to obtain the corresponding discrete curvature field matrix κ(i,j).

[0020] The resistance-curvature mapping algorithm further comprises a mapping function construction step performed before the resistance-curvature real-time mapping step, which comprises the following:

[0021] According to the initial resistance matrix R_flat of the integrated flexible sensing film in the unstressed state, the resistance relative change rate of each array unit is calculated as ε r (i,j)=(R_k(i,j)-R_flat(i,j)) / R_flat(i,j), and according to offline experimental calibration, a series of data points (ε r ,κ) are obtained.

[0022] Polynomial fitting is used to perform regression analysis on these data, and the mapping function f is obtained as κ=f(ε r )=α·ε r +β·ε r 2+..., where κ is the curvature, and α and β are calibration coefficients obtained by fitting.

[0023] The surface reconstruction algorithm comprises the following:

[0024] The discrete curvature field matrix κ(i,j) is expanded into a continuous curvature function κ(x,y) defined on the entire antenna array region (x,y) through B-spline interpolation;

[0025] The continuous curvature function κ(x,y) is integrated twice, and the boundary conditions are set to inversely solve the final height function z(x,y).

[0026] The microwave imaging inversion algorithm specifically comprises the following:

[0027] For each antenna unit n, its real coordinates in the three-dimensional space are calculated as z_n=z(x_n,y_n) P_n_corrected=(x_n,y_n,z_n) using its nominal coordinates (x_n,y_n) and the height field function z(x,y).

[0028] The corrected electromagnetic wave propagation path parameter is calculated according to d'_ij = ||P_i_corrected - P_j_corrected||;

[0029] A matrix describing the system response is constructed, and the system matrix is recalculated and updated using the corrected three-dimensional coordinates {P_n_corrected}, to ensure that the mathematical model of the imaging inversion is completely consistent with the physical reality of the device.

[0030] The present application has the following advantages: a wearable breast cancer intelligent screening system realizes real-time perception and adaptive correction of the antenna array geometric posture by deeply fusing flexible sensing technology and microwave imaging technology, and ensures the accuracy and repeatability of the detection results. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 The figure is a structural schematic diagram of the present application;

[0032] Figure 2 The figure is a structural schematic diagram of the integrated flexible sensing film;

[0033] Figure 3 The figure is a structural schematic diagram of the antenna path switching module. DETAILED DESCRIPTION

[0034] To make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in combination with the drawings herein is not intended to limit the protection scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application. The present application will be further described below in combination with the drawings.

[0035] The present application specifically relates to a wearable breast cancer intelligent screening system, aiming at the problems of insufficient imaging accuracy of existing wearable breast detection devices and large volume and non-wearable of fixed microwave imaging systems, and considering the problems of array geometric distortion, lack of posture perception, data asynchronization and limited acquisition efficiency in flexible wearable microwave devices, the present application realizes the closed-loop optimization of "shape perception, real-time correction and accurate imaging" through the deep fusion of flexible sensing and microwave imaging, which is different from simple hardware integration, but realizes the geometric self-adaptation through posture driving.

[0036] AsFigure 1 As shown, it includes a wearable detection unit and a master processing unit, the two parts are connected through a radio frequency cable and a control cable, the master processing unit is connected with a PC (terminal computer) to realize the separation of signal transmission and system control, which not only ensures the lightweight of the wearable part, but also ensures the measurement accuracy and processing capacity.

[0037] Among them, the wearable detection unit is configured to sense the resistance change when the external force or pressure acts through the integrated flexible sensing film, and send the resistance calculated by the sampling chip to the master processing unit after collecting the voltage;

[0038] The master processing unit is configured to perform the following functions: a) obtaining the resistance data sent by the wearable detection unit, which represents the geometric pose thereof; b) controlling the antenna path switching module to select a transmitting antenna and a receiving antenna pair in the antenna array, and measuring the S21 parameter between the antenna pair by the network analysis module; c) by controlling the switch sequence, realizing the multi-path polling scanning of different sensor unit resistance data and different antenna pair S21 parameters;

[0039] The PC is configured to convert the obtained data into a curvature field matrix describing the bending degree of the antenna array through a resistance-curvature mapping algorithm, and obtain the geometric pose data describing the three-dimensional space form of the antenna array at the moment of wearing through a surface reconstruction algorithm, correct the antenna coordinate parameters in the microwave imaging inversion algorithm through the geometric pose data, generate the breast internal dielectric property image eliminating the deformation error, and screen the breast cancer according to the image.

[0040] Further, the wearable detection unit specifically includes two detection parts of left and right sides arranged symmetrically, respectively used for covering the left and right breasts. Each side detection part is integrated with a set of independent detection hardware, including a flexible antenna, an integrated flexible sensing film, two array gating chips and a sampling chip. In each side detection part, the integrated flexible sensing film includes a plurality of array units arranged in an array, which is arranged at the position corresponding to the breast in the lining of the bra. One of the cores of the present application lies in the design of the antenna array. In each side detection part, a 4x4 flexible antenna array composed of 16 antenna units is included, so the whole wearable system includes 32 antenna units. This bilateral and multi-unit design ensures comprehensive signal coverage of bilateral breast tissue. Each antenna unit adopts a coplanar waveguide structure, has low profile, wide frequency band and good bending performance; the working frequency covers 10MHz-6GHz, the array adopts a cross arrangement, every four antenna units form a group, and is distributed in a cross shape, so as to improve the view coverage rate and spatial resolution.

[0041] The two array gating chips are connected with each row of flexible antennas and each column of flexible antennas respectively, and the two array gating chips are connected with the sampling chip.

[0042] Further, as shown in Figure 2 The integrated flexible sensing film includes a protective layer made of polydimethylsiloxane (PDMS) material on the upper and lower layers and an AGNW layer in the middle; the AgNW layer is deposited by spraying and annealing, and is then precisely patterned into a 4x4 or 8x8 square array unit (unit size 5mmx5mm) by laser ablation; the upper layer of PDMS is used to protect the conductive electrode and the wiring to prevent external environmental interference, and the lower layer of PDMS serves as a base layer to ensure that the sensing film can be stably attached to the antenna array and the surface of the breast; the thickness of the two layers ranges from 200μm to 500μm.

[0043] Each array unit is a resistive sensor, and when subjected to external force bending or pressure, the contact resistance of the AgNW layer changes reversibly; when the film is bent or stretched, the contact points between the silver nanowires decrease, and the nanowire network structure is rearranged, resulting in an increase in contact resistance; when the film is compressed, the nanowire contacts are more closely spaced, the contact area increases, and the contact resistance decreases; through row and column line selection, for a 4x4 array, four row lines and four column lines are used for matrix selection.

[0044] The row and column line selection is realized by two TMUX1108PWR chips, which can realize fast switching of 8-channel analog signals and have a low on-resistance; each array unit is sampled by an ADS1256 sampling chip to calculate the voltage and resistance, and according to the circuit shown in the figure, the resistance R=(V + -V - ) / I can be calculated, where I=100uA is the constant current output by the REF200 chip; the row and column electrodes are collected through narrow conductive wiring to the square fan-out area at the edge of the film, and are connected to the external circuit board through FPC ribbon.

[0045] Further, the main control processing unit includes an antenna path switching module, a network analysis module, an MCU, and a power module; the wearable detection unit is connected to the antenna path switching module, and the switching sequence of the antenna path switching module is controlled by the MCU to realize dynamic switching between the wearable detection unit and the main control processing unit, and to realize multi-path polling scanning; the network analysis module is connected to the antenna path switching module and the MCU, and is used to measure the S21 parameters transmitted under different channel combinations; the power module provides power supply for the MCU, the antenna path switching module, and the network analysis module.

[0046] As shown in Figure 3As shown, in order to efficiently manage the signal transmission and reception of 32 antenna units on both sides, the antenna path switching module in the main control processing unit adopts a hierarchical switching design. The module includes two main input / output ports (TX_IN and RX_IN) connected to an external vector network analyzer (VNA). The signal path first passes through a double-pole double-throw switch (DPDT) to select the left or right antenna array branch, thereby achieving scanning of a single breast. After selecting a side branch, two 16-channel RF switches (TX-SP16T and RX-SP16T) are used to accurately distribute the transmit (TX) and receive (RX) signals to the target channels in the 16 antenna units on that side.

[0047] The network analysis module is composed of a two-port vector network analyzer (VNA) or a lightweight RF acquisition module, which is used to measure the transmission parameters S21 under different channel combinations, including amplitude and phase. The TX port is connected to TX_IN of the path switching module, and the RX port is connected to RX_IN. The data is uploaded to the PC through a high-speed interface (USB / optical fiber) for subsequent processing.

[0048] Further, the goal of the resistance-curvature mapping algorithm is to establish an accurate mathematical relationship between the resistance change of each sensing unit on the flexible sensing film and the local curvature of its location. That is, the collected multi-point, discrete resistance value matrix is converted into a curvature field matrix that also describes the degree of geometric bending. This is the first step in realizing the transition from electrical measurement to geometric reconstruction, which includes the following steps:

[0049] At a certain channel signal acquisition moment, the resistance value t_k, the posture data matrix R_k collected and packaged by the MCU, and each element R_k(i,j) in the matrix represents the instantaneous resistance value of the (i,j) sensing unit;

[0050] The posture data matrix R_k is input into the mapping function f to obtain the corresponding discrete curvature field matrix κ(i,j).

[0051] Further, the resistance-curvature mapping algorithm also includes a mapping function construction step, which is executed before the resistance-curvature real-time mapping step, and includes the following content:

[0052] First, the system needs a reference state, i.e., the initial resistance matrix R_flat of the sensing film in the completely flat, unforced state. This data is calibrated and stored before the device is shipped, and the relative resistance change rate of each array unit is calculated as ε r (i,j) = (R_k(i,j) - R_flat(i,j)) / R_flat(i,j), and according to the offline experimental calibration, a series of data points (ε r, κ), i and j represent the i-th row and j-th column, R_k(i,j) represents the resistance of the array unit in the i-th row and j-th column after bending;

[0053] Further, the offline experimental calibration process aims to establish the mapping relationship between the relative change rate ε r and the local curvature κ, and the specific steps can include: tightly attaching the integrated flexible sensing film to a series of physical molds with precise known curvatures (for example, cylinders with different radii or specific curved surfaces printed by three-dimensional printing), and synchronously recording the resistance values of each sensing unit under different known curvatures, so as to obtain a series of data point pairs (ε r , κ). In addition, to further improve the accuracy, the calibration process can also introduce a pressure sensor for cross verification, so as to establish a more complex multivariate mapping function capable of decoupling the common influence of normal pressure and bending deformation on the resistance, thereby more accurately reflecting the real geometric posture of the antenna array.

[0054] The mapping function f is obtained by using polynomial fitting to perform regression analysis on the data, that is, κ = f(ε r ) = α·ε r + β·ε r 2 +..., where κ is the curvature, and α and β are calibration coefficients obtained by fitting.

[0055] At this point, the resistance matrix without physical meaning has been converted into key geometric information describing the local bending degree of the antenna array.

[0056] Further, the goal of the surface reconstruction algorithm is to use the discrete curvature field matrix κ(i,j) output above to reconstruct a continuous and smooth three-dimensional surface of the entire flexible antenna array region through mathematical methods, including the following contents:

[0057] The 4x4 curvature sampling points are discrete. In order to obtain the information of the entire surface, interpolation is needed. The preferred method of the present application is bicubic B-spline interpolation. The reason for choosing B-spline interpolation is that it can generate a smooth surface with continuous second-order derivatives (C 2 2 continuity), which is highly consistent with the natural bending form of the flexible substrate in the physical world, avoiding the sharp creases that may be generated by linear interpolation and the like.

[0058] Through B-spline interpolation, the discrete curvature field matrix κ(i,j) is expanded to a continuous curvature function κ(x,y) defined on the entire antenna array region (x,y);

[0059] In a preferred embodiment of the present application, when the deformation amplitude is small, an efficient algorithm based on small deflection deformation assumption can be used. Under this assumption, the curvature of the surface in X and Y directions is approximately proportional to the second order partial derivative of the height function z(x, y): Therefore, reconstructing the surface is equivalent to solving a Poisson's equation, and the height field function z(x, y) can be inversely solved by integrating the continuous curvature function K(x, y) twice. To ensure the uniqueness of the solution, boundary conditions need to be introduced. For example, it can be assumed that the edge area of the bra has small deformation, and its height z or slope is zero. These constraints are introduced into the solution process of the partial differential equation (e.g. numerical solution using finite difference method or finite element method). It should be understood by those skilled in the art that the scope of protection of the present application is not limited to this. For scenarios with large deformation or higher accuracy, more advanced surface reconstruction algorithms such as nonlinear finite element analysis (FEM) can also be used to solve the height field function, and these alternatives are within the scope of protection of the present application.

[0060] After solving, the final height field function z(x, y) is obtained. This function accurately describes the three-dimensional spatial form of the antenna array at the moment of wearing in mathematical form.

[0061] Further, the microwave imaging inversion algorithm aims to apply the three-dimensional surface information reconstructed in the previous module to the microwave imaging inversion algorithm to make real-time corrections to the key parameters that depend on the antenna geometry layout, including the following:

[0062] For each antenna element n, its height field coordinate in three-dimensional space is calculated as z_n = z(x_n, y_n) using its nominal coordinates (x_n, y_n) and height field function z(x, y), and its actual spatial coordinates are P_n_corrected = (x_n, y_n, z_n), where n = 1, 2,..., 16, completing the mapping from ideal planar layout to real spatial layout;

[0063] Correcting electromagnetic wave propagation path parameters: The core of microwave imaging algorithms (such as delay-and-sum algorithm, DBIM) is to accurately calculate the propagation delay and path loss between the transmitting antenna and the receiving antenna, and between the antenna and each pixel point in the imaging area. These calculations are extremely dependent on the Euclidean distance between the antennas. The corrected electromagnetic wave propagation path parameters are calculated according to d'_ab=||P_a_corrected-P_b_corrected||, which provides accurate geometric input for subsequent calculations. P_a_corrected represents the three-dimensional spatial coordinates of the a-th antenna unit after deformation of the integrated flexible sensing film, P_b_corrected represents the three-dimensional spatial coordinates of the b-th antenna unit after deformation of the integrated flexible sensing film, and d'_ab represents the distance between the two antenna units.

[0064] For more complex iterative imaging algorithms (such as contrast source inversion method), the algorithm constructs a Jacobian matrix or Green's function matrix describing the system response. Each item of the matrix is related to the antenna position and its propagation function to the imaging area. The system matrix is recalculated and updated using the corrected three-dimensional coordinates {P_n_corrected}, ensuring that the mathematical model of the imaging inversion is completely consistent with the physical reality of the device.

[0065] At this point, the S21 parameters with bound attitude information can be input into the corrected imaging algorithm, and finally a high-precision, deformation error-eliminated dielectric property image of the breast is generated.

[0066] The above description is only the preferred embodiment of the present application, and it should be understood that the present application is not limited to the form disclosed herein, and should not be considered as excluding other embodiments, but can be used in various other combinations, modifications and improvements, and can be modified within the scope of the concept described herein, by the above teaching or related technical or knowledge. The modifications and changes made by those skilled in the art without departing from the spirit and scope of the present application shall be within the scope of protection of the appended claims of the present application.

Claims

1. A wearable intelligent breast cancer screening system, characterized in that: It includes a wearable detection unit with a dual-sided symmetrical design and a main control processing unit connected to it, which is connected to a PC; The wearable detection unit is configured to sense the resistance change when an external force or pressure is applied by an integrated flexible sensing film, and to collect the voltage through a sampling chip, calculate the resistance, and send it to the main control processing unit. The main control processing unit is configured to perform the following functions: a) acquire resistance data representing the geometric orientation of the wearable detection unit; b) control the antenna path switching module to select the transmitting antenna and receiving antenna pair in the antenna array, and have the network analysis module measure the S21 parameters between the antenna pairs; c) achieve multi-path polling scan of resistance data of different sensing units and S21 parameters of different antenna pairs by controlling the switching sequence. The PC is configured to convert the acquired data into a curvature field matrix describing the degree of bending of the antenna array through a resistance-curvature mapping algorithm, and then obtain geometric attitude data describing the three-dimensional spatial morphology of the antenna array at the moment of wearing through a surface reconstruction algorithm. The antenna coordinate parameters in the microwave imaging inversion algorithm are corrected through the geometric attitude data, and an image of the internal dielectric properties of the breast is generated to eliminate deformation errors. Breast cancer is then screened based on the image.

2. The wearable intelligent breast cancer screening system according to claim 1, characterized in that: The wearable detection unit includes two symmetrically arranged detection components, left and right, which are used to fit the bilateral breasts respectively. Each detection component includes a flexible antenna, an integrated flexible sensing film, two array gating chips and a sampling chip. The integrated flexible sensing film includes multiple array units arranged in an array, which are disposed at the position of the bra that fits the breast. Flexible antennas are arranged in an array and disposed between each row and each column of array units. Each row of flexible antennas is connected to each row of array units, and each column of flexible antennas is connected to each column of array units. Two array gating chips are connected to each row of flexible antennas and each column of flexible antennas, respectively, and two array gating chips are connected to the sampling chip.

3. The wearable intelligent breast cancer screening system according to claim 2, characterized in that: The integrated flexible sensing film includes a protective layer made of polydimethylsiloxane material on both the top and bottom and an AGNW layer in the middle; the AgNW layer is formed by spraying, deposition and annealing, and then precisely patterned into an n×n square array unit by laser ablation; Each array unit is a resistive sensor. When subjected to external bending or pressure, the contact resistance of the AgNW layer undergoes a reversible change. When the film is bent or stretched, the contact points between the silver nanowires decrease, and the nanowire network structure rearranges, leading to an increase in contact resistance. When the film is compressed, the nanowires are in closer contact, the contact area increases, and the contact resistance decreases. The rapid switching of multi-channel analog signals is achieved through two array gating chips. Each array unit collects voltage and calculates resistance through a sampling chip and sends the data to the main control processing unit.

4. The wearable intelligent breast cancer screening system according to claim 1, characterized in that: The main control processing unit includes an antenna path switching module, a network analysis module, an MCU, and a power supply module. The wearable detection unit is connected to the antenna path switching module. The MCU controls the switching sequence of the antenna path switching module to achieve dynamic switching between the wearable detection unit and the main control processing unit, thereby realizing multi-path polling scanning. The network analysis module is connected to the antenna path switching module and the MCU and is used to measure the S21 parameters transmitted under different channel combinations. The power supply module provides power to the MCU, the antenna path switching module, and the network analysis module.

5. The wearable intelligent breast cancer screening system according to claim 1, characterized in that: The resistance-curvature mapping algorithm includes a real-time resistance curvature mapping step, which includes the following: The instantaneous resistance value t_k of a certain channel signal acquisition is collected by the MCU and packaged into an attitude data matrix R_k. Each element R_k(i,j) in the matrix represents the instantaneous resistance value of the (i,j)th sensing unit, where i and j represent the i-th row and j-th column, respectively. R_k(i,j) represents the resistance of the array unit in the i-th row and j-th column after bending. The attitude data matrix R_k is input into the mapping function f to obtain the corresponding discrete curvature field matrix κ(i,j).

6. A wearable intelligent breast cancer screening system according to claim 5, characterized in that: The resistance-curvature mapping algorithm further includes a mapping function construction step, which is executed before the real-time resistance-curvature mapping step, and includes the following: Based on the initial resistance matrix R_flat of the integrated flexible sensing film under unstressed conditions, the relative resistance change rate ε of each array unit is calculated. r (i,j)=(R_k(i,j)-R_flat(i,j)) / R_flat(i,j), and based on offline experimental calibration, a series of data point pairs (ε) are obtained. r ,κ); Polynomial fitting was used to perform regression analysis on these data, and the mapping function f was obtained as κ=f(ε) r )=α·ε r +β·ε r 2+..., where κ is the curvature, and α and β are calibration coefficients obtained through fitting.

7. A wearable intelligent breast cancer screening system according to claim 5, characterized in that: The surface reconstruction algorithm includes the following: By using B-spline interpolation, the discrete curvature field matrix κ(i,j) is extended into a continuous curvature function κ(x,y) defined over the entire antenna array region (x,y); By integrating the continuous curvature function κ(x,y) twice and setting boundary conditions, the final height function z(x,y) can be obtained by inverse solving.

8. A wearable intelligent breast cancer screening system according to claim 5, characterized in that: The microwave imaging inversion algorithm specifically includes the following: For each antenna element n, its nominal coordinates (x_n, y_n) and height field function z(x, y) are used to calculate its true coordinates in three-dimensional space as z_n = z(x_n, y_n)P_n_corrected = (x_n, y_n, z_n); The corrected electromagnetic wave propagation path parameters are calculated based on d'_ij=||P_i_corrected-P_j_corrected||, where P_a_corrected represents the three-dimensional spatial coordinates of the a-th antenna element after deformation of the integrated flexible sensing film, P_b_corrected represents the three-dimensional spatial coordinates of the b-th antenna element after deformation of the integrated flexible sensing film, and d'_ab represents the distance between the two antenna elements. Construct a matrix describing the system response, and recalculate and update this system matrix using the corrected 3D coordinates {P_n_corrected} to ensure that the mathematical model of the imaging inversion is completely consistent with the physical reality of the device.

Images