Bra for early breast cancer detection based on AI analysis and detection method thereof

Abstract

The invention provides a bra for early breast cancer detection based on AI analysis and a detection method thereof, and relates to the field of medical instruments, and the bra comprises a signal transmitting module used for transmitting low-voltage high-frequency alternating current and outputting the low-voltage high-frequency alternating current to an EIT sensor; the bra main body is used for collecting electrical impedance information of mammary tissue; the signal analysis module is used for extracting electrical impedance characteristic parameters related to the physiological state of the breast tissue from the preprocessed digital signal; the communication transmission device is used for coding and modulating the electrical impedance characteristic parameters into wireless signals and transmitting the wireless signals to a mobile phone end, and managing power supply and electric quantity monitoring at the same time; and the AI data analysis software is used for identifying abnormal conditions of the mammary tissue based on a deep learning algorithm and generating a visual report according to an identification result. The brassiere is combined with the EIT sensor array, so that the electrical impedance information of the mammary tissue can be acquired in real time, the physiological state change of the mammary tissue is reflected, and the early discovery of mammary gland abnormality is facilitated.

Description

Technical Field

[0001] The present invention relates to the field of medical devices. Specifically, it particularly relates to a bra for early breast cancer detection based on AI analysis and its detection method. Background Art

[0002] As one of the most common malignant tumors in women, the incidence rate of breast cancer shows an upward trend globally. Timely detection of early breast cancer is crucial for improving the cure rate and the prognosis of patients. Currently, the detection methods for breast cancer mainly include self-examination, clinical examination, imaging examinations (such as mammography, ultrasound examination, MRI, etc.), and cytological examination. However, these traditional detection means still have many limitations.

[0003] Currently, there are many limitations in the existing breast detection methods. Although traditional mammography is a commonly used screening method, its diagnostic accuracy for dense breast tissue is relatively low, and it is easy to miss a diagnosis. Moreover, the breast will be squeezed to a certain extent during the examination, causing discomfort to the patient. Although ultrasound examination is relatively simple to operate, its results are greatly affected by the operator's experience, and the detection accuracy is unstable. Magnetic resonance imaging (MRI) examination has relatively high accuracy, but the equipment is expensive, the examination cost is high, the examination time is long, and it requires injection of a contrast agent and has certain radioactivity, which limits its wide application.

[0004] Regarding the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] In view of this, the present invention provides a bra for early breast cancer detection based on AI analysis and its detection method to solve the above-mentioned problems.

[0006] To solve the above problems, the specific technical solutions adopted by the present invention are as follows: According to one aspect of the present invention, there is provided a bra for early breast cancer detection based on AI analysis, including: A signal emission module, configured to control a signal emission module located on the bra main body to emit a low-voltage high-frequency alternating current output to the EIT sensor through a mobile phone terminal; The bra main body, configured to collect breast tissue impedance information through the EIT sensor; A signal analysis module, configured to preprocess the breast tissue impedance information and extract impedance characteristic parameters related to the physiological state of the breast tissue from the preprocessed digital signal; A communication transmission device, configured to encode and modulate the impedance characteristic parameters into a wireless signal and transmit it to an external receiving device, and manage power supply and power monitoring at the same time; An AI data analysis software, which is used to formulate a signal emission plan, receive impedance characteristic parameters, identify abnormal conditions of breast tissue based on a deep learning algorithm, and generate a visualization report according to the identification result.

[0007] Preferably, the bra body is made of a flexible and breathable medical-grade elastic fabric, and an installation cavity is provided inside the bra body. A number of EIT sensor arrays are arranged inside the installation cavity. Shoulder straps are also arranged on both sides of the bra body, and a power supply module is arranged on the outer side of one of the shoulder straps. One side of the power supply module is a signal emission module, and the signal emission module is connected to one of the shoulder straps; The signal emission module is used to receive a control instruction from the mobile phone terminal and emit a low-voltage high-frequency alternating current signal.

[0008] Preferably, the signal emission module includes: A signal receiving unit, which is used to receive the signal emission plan formulated by the AI data analysis software on the mobile phone terminal; A signal emitting unit, which is used to emit high-frequency low-voltage alternating current to the EIT sensor array regularly and quantitatively according to the signal emission plan; Preferably, the signal analysis module includes: A signal preprocessing unit, which is electrically connected to the EIT sensor array, receives and preprocesses the impedance information. The preprocessing includes filtering, amplification, and analog-to-digital conversion; A feature extraction unit, which is used to extract impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information.

[0009] Preferably, the preprocessing of the impedance information includes: Based on wavelet transform and Wiener filtering method, filter the impedance information; Use a low-noise amplifier to amplify the filtered impedance information; Use a mode converter to convert the amplified impedance information into a digital signal.

[0010] Preferably, the extraction of impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information includes: Perform a first-level resolution discrete Fourier transform on the preprocessed impedance signal, and calculate the center frequency and determine the frequency range where the power drops to half of the peak value according to the discrete Fourier transform result; Based on the center frequency and the frequency range where the power drops to half of the peak value, optimize the discrete Fourier transform parameters and perform a second-level resolution discrete Fourier transform on the impedance signal; Extract impedance characteristics related to the physiological state of breast tissue from the high-resolution DFT result.

[0011] Preferably, the AI data analysis software includes: A signal emission control module, configured to formulate a signal emission plan and transmit it to the signal emission module on the bra body; A data transmission and reception module, configured to transmit a signal control instruction through a communication transmission device, receive the impedance characteristic parameters sent, and temporarily store them; An AI tumor recognition module, configured to input the impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judge breast tissue abnormality, analyze the type and degree of abnormality according to the output result of the AI tumor recognition model, and analyze the impedance characteristic parameters to restore the original breast detection result information; A visualization display module, configured to display the original breast detection result information on the software interface and present it to the user in an intuitive chart, image or text form; A report generation module, configured to generate a standardized or personalized breast detection report according to the user's selection.

[0012] Preferably, inputting the impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judging breast tissue abnormality, analyzing the type and degree of abnormality according to the output result of the AI tumor recognition model, and analyzing the impedance characteristic parameters to restore the original breast detection result information includes: Collecting breast case data, constructing and training a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model; Taking the impedance characteristic parameters as input, using the AI tumor recognition model to infer and output the corresponding breast tissue classification result; Judging breast tissue abnormality and analyzing the type and degree of abnormality according to the result output by the AI tumor recognition model; Analyzing the combined impedance characteristic parameters, and restoring the complete breast detection result according to the judgment and analysis results.

[0013] Preferably, collecting breast case data, constructing and training a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model includes: Collecting and processing breast case data, where the breast case data includes impedance signals and case labels, and the data processing includes: standardizing, denoising and feature extracting the impedance signals; Constructing a neural network structure, using the processed case data, and iteratively training the neural network structure by adopting the S-BGD and gradient accumulation strategies; Updating the neural network parameters through the backpropagation algorithm to obtain a trained neural network model and using it as the AI tumor recognition model.

[0014] Preferably, a neural network structure is constructed, and the processed case data is used, and the neural network structure is iteratively trained by using the S-BGD and gradient accumulation strategies, including: Construct the input layer, hidden layer, and output layer of the neural network structure; Use the processed case data as the training set, and preliminarily train the neural network structure by using the stochastic gradient descent method; Obtain the training rounds of the neural network structure, and when the training rounds reach the preset threshold, use batch gradient descent to optimize and train the neural network structure until the training ends.

[0015] According to another aspect of the present invention, a detection method for early breast cancer detection based on AI analysis is provided, including the following steps: S1. Control the signal emission module on the bra main body through the mobile phone to emit low-voltage high-frequency alternating current and output it to the EIT sensor; S2. Collect the impedance information of breast tissue through the EIT sensor; S3. Preprocess the impedance information of breast tissue, and extract impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed digital signal; S4. Encode and modulate the impedance characteristic parameters into wireless signals and transmit them to an external receiving device, and manage power supply and power monitoring at the same time; S5. Develop a signal emission plan, receive the impedance characteristic parameters, and identify abnormal conditions of breast tissue based on a deep learning algorithm, and generate a visual report according to the identification result.

[0016] The beneficial effects of the present invention are as follows: 1. The bra of the present invention combines an EIT sensor array, can collect the impedance information of breast tissue in real time, reflect the changes in the physiological state of breast tissue, and helps to detect breast abnormalities at an early stage.

[0017] 2. The signal analysis module of the present invention efficiently analyzes the collected data, improving the accuracy and reliability of the impedance signals of breast diseases.

[0018] 3. The communication transmission device of the present invention realizes the wireless transmission of breast detection results, facilitating patients to transmit data to the AI analysis software in a timely manner for further analysis and diagnosis.

[0019] 4. The AI data analysis software of the present invention accurately identifies abnormal conditions of breast tissue through a tumor recognition model, displays the detection results in an intuitive manner, and generates a detailed report, providing a comprehensive reference for doctors' diagnosis and also facilitating patients to understand their own breast health status. Brief Description of the Drawings

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings. In the drawings: Figure 1 is a schematic block diagram of a bra for early breast cancer detection based on AI analysis according to an embodiment of the present invention; Figure 2 is a schematic structural diagram of the bra body in a bra for early breast cancer detection based on AI analysis according to an embodiment of the present invention; Figure 3 is a flowchart of a detection method for early breast cancer detection based on AI analysis according to an embodiment of the present invention.

[0021] In the figure: 1. Signal emission module; 2. Bra body; 201. Installation cavity; 202. EIT sensor array; 203. Shoulder strap; 204. Power supply module; 205. Signal emission module; 3. Signal analysis module; 4. Communication transmission device; 5. AI data analysis software. Detailed implementation manners

[0022] In order to enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0023] According to an embodiment of the present invention, a bra for early breast cancer detection based on AI analysis and its detection method are provided.

[0024] Now, the present invention will be further described in combination with the accompanying drawings and specific implementation manners. As Figure 1 shown, according to an embodiment of the present invention, a bra for early breast cancer detection based on AI analysis is provided, including: A signal emission module 1, configured to control a signal emission module 205 located on the bra body 2 through a mobile phone terminal to emit a low-voltage high-frequency alternating current and output it to the EIT sensor; Specifically, a control instruction is sent using the AI data analysis software on the mobile phone side. The signal transmission module 205 on the bra main body 2 generates a high-frequency alternating current signal, and the high-frequency alternating current signal acts on the breast tissue through the EIT sensor. The EIT sensor array receives the impedance response of the breast tissue to the alternating current signal.

[0025] The bra main body 2 is used to collect the impedance information of the breast tissue through the EIT sensor; As a preferred embodiment, as Figure 2 shown, the bra main body is made of a flexible and breathable medical-grade elastic fabric, and an installation cavity is provided on the inner side of the bra main body. A number of EIT sensor arrays are arranged inside the installation cavity. Shoulder straps 203 are also provided on both sides of the bra main body 2, and a power supply module 204 is provided on the outer side of one of the shoulder straps 203 for providing power. One side of the power supply module 204 is a signal transmission module 205, and the signal transmission module 205 is connected to one of the shoulder straps 203; a signal analysis module 3 and a communication transmission device 4 are provided on the outer side of the other shoulder strap 203.

[0026] Specifically, the signal transmission module 205 includes: A signal receiving unit for receiving the signal transmission plan formulated by the AI data analysis software on the mobile phone side; A signal transmitting unit for regularly and quantitatively transmitting high-frequency low-voltage alternating current to the EIT sensor array according to the signal transmission plan.

[0027] It should be noted that the shape of the installation cavity 201 is adapted to the contour of the female breast to ensure close fit during detection. A plurality of EIT sensor arrays 202 are provided at positions corresponding to the left and right quadrants of the breast, evenly distributed at a spacing of 5 mm in the front, back, up, and down directions and seamlessly connected to the inner wall of the bra main body 1 to collect the impedance information of the breast tissue. The sensors are designed with high precision and low noise to effectively suppress external electromagnetic interference.

[0028] In addition, the electrodes of each EIT sensor array 202 are distributed in a ring shape from front to back, divided into four regions of up, down, left, and right. The left and upper regions are the emission regions, and the right and lower regions are the reception regions. The electrode intervals are 3 - 5 mm both left and right and front and back. The communication lines of each group of electrodes are connected to the battery and signal processing and transmission unit at the back through the left and right bra straps. The battery / signal transmission unit is designed with a detachable TYPC port for easy cleaning.

[0029] The signal analysis module 3 is used to preprocess the impedance information of the breast tissue and extract impedance characteristic parameters related to the physiological state of the breast tissue from the preprocessed digital signal; As a preferred embodiment, the signal analysis module 3 includes: A signal preprocessing unit, which is electrically connected to the EIT sensor array 202, receives and preprocesses the impedance information, and the preprocessing includes filtering, amplification, and analog-to-digital conversion; As a preferred embodiment, preprocessing the impedance information includes: Filtering the impedance information based on wavelet transform and Wiener filtering method; Amplifying the impedance information after filtering using a low-noise amplifier; Converting the impedance information after amplification into a digital signal using a mode converter.

[0030] It should be noted that wavelet transform can effectively decompose the signal into different frequency sub-bands. Based on the pre-selected wavelet basis function and decomposition level, the noise and useful signals in the impedance information can be separated into different sub-bands. The specific implementation is as follows: First, according to the characteristics of the impedance signal, the Daubechies wavelet is used as the wavelet basis function; then, based on the frequency range of the impedance signal and the distribution of the noise, the decomposition level of the wavelet is determined; then, according to the wavelet decomposition level, the impedance signal is decomposed by wavelet to obtain the wavelet coefficients on different frequency sub-bands; according to the obtained wavelet coefficients, the wavelet coefficients on each frequency sub-band are thresholded using the hard threshold method or the soft threshold method to remove the wavelet coefficients corresponding to the noise; finally, wavelet reconstruction is performed based on the wavelet coefficients to obtain the filtered impedance signal.

[0031] A feature extraction unit, which is used to extract impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information.

[0032] As a preferred embodiment, extracting impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information includes: Performing a first-level resolution discrete Fourier transform on the preprocessed impedance signal, and calculating the center frequency and determining the frequency range where the power drops to half of the peak value according to the discrete Fourier transform result; It should be noted that the Fourier transform of the first-level resolution is a preliminary frequency-domain analysis of the signal, which can identify the main frequency components of the impedance signal.

[0033] The center frequency refers to the frequency that best represents the characteristics of the signal in the signal spectrum. It is calculated by the method of weighted average, and the calculation formula of the center frequency is: ; In the formula, f i represents the i-th frequency point, P(f i ) represents the power value of this frequency point, and f c represents the center frequency.

[0034] It should be understood that the frequency range where the power drops to half of the peak represents the main frequency band width of the signal and is used to describe the signal bandwidth. Based on the center frequency and the frequency range where the power drops to half of the peak, by optimizing the discrete Fourier transform parameters, a two-level resolution discrete Fourier transform is performed on the impedance signal; Specifically, based on the calculation results of the center frequency and the power bandwidth range, the parameters of the discrete Fourier transform, such as the sampling rate and the frequency resolution, can be further optimized.

[0035] On the basis of the first-level transform, through a higher-resolution Fourier transform (i.e., the two-level resolution Fourier transform), the frequency resolution can be improved to more clearly observe the detailed features of the impedance signal. In particular, the minute changes in the high-frequency part can more accurately reveal the physiological state of the breast tissue.

[0036] It should be noted that impedance characteristics closely related to the physiological state of the breast tissue can be extracted from the two-level resolution Fourier transform results. For example, factors such as the growth of breast tumors, tissue heterogeneity, or local blood flow changes will have specific effects on the impedance signal. These effects are manifested as changes in certain specific frequency bands in the frequency domain. Common characteristic parameters include: amplitude spectrum, phase spectrum, and bandwidth and spectral smoothness; The communication transmission device 4 is used to encode and modulate the impedance characteristic parameters into a wireless signal and transmit it to an external receiving device, while managing the power supply and power monitoring; Specifically, the characteristic parameter eigenvalues are usually digital data. Before transmission, they need to be converted into a format suitable for wireless transmission through an encoder. Common coding methods include: binary coding and quantization coding; In addition, in order to make the bra portable, optimization needs to be carried out in terms of battery life and power consumption. For example: (1) Equipping with a high-capacity and long-life battery (such as a lithium battery) can enable the battery to work continuously without frequent charging; (2) Integrating a power management chip (such as chips provided by manufacturers such as TI and Maxim) can efficiently manage the charging and discharging of the battery, monitor the battery health, and extend the battery life; (3) Configuring a power monitoring module, such as using a current sensor and a voltage sensor, can real-time track the voltage, current, and remaining power of the battery, and real-time collect the working state data of the battery.

[0037] The AI data analysis software 5 is used to receive the impedance characteristic parameters and identify abnormal conditions of the breast tissue based on deep learning algorithms, and generate a visualization report according to the identification results.

[0038] As a preferred implementation, the AI data analysis software 5 includes: A signal transmission control module, which is used to formulate a signal transmission plan and transmit it to the signal transmission module 205 on the bra main body 2; A data transmitting and receiving module, which is used to transmit signal control instructions through a communication transmission device 4, receive the impedance characteristic parameters sent, and perform temporary storage; An AI tumor recognition module, which is used to input the impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judge the abnormality of breast tissue, analyze the type and degree of abnormality according to the output result of the AI tumor recognition model, and analyze the impedance characteristic parameters to restore the original breast detection result information; As a preferred implementation manner, input the impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judge the abnormality of breast tissue, analyze the type and degree of abnormality according to the output result of the AI tumor recognition model, and analyze the impedance characteristic parameters to restore the original breast detection result information, including: Collect breast case data, construct and train a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model; As a preferred implementation manner, collecting breast case data, constructing and training a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model includes: Collect breast case data and perform data processing. The breast case data includes impedance signals and case labels. The data processing includes: normalizing, denoising, and feature extraction of the impedance signals; Construct a neural network structure, use the processed case data, and adopt the S-BGD and gradient accumulation strategies to iteratively train the neural network structure; It should be noted that the impedance signal is obtained by measuring the impedance characteristics of breast tissue. The impedance signal can reflect the physiological and pathological states of breast tissue. Breast tissue exhibits different impedance characteristics under different physiological or pathological states. For example, there are obvious differences in the impedance values between healthy tissue and tumor tissue.

[0039] Case labels are usually given by doctors according to standard medical detection methods (such as pathological biopsy) and label the specific diagnosis results of each case, including whether it is a tumor, tumor type (benign or malignant), tumor size, shape, etc.

[0040] As a preferred implementation manner, constructing a neural network structure, using the processed case data, and adopting the S-BGD and gradient accumulation strategies to iteratively train the neural network structure includes: Construct the input layer, hidden layer, and output layer of the neural network structure; Specifically, the number of nodes in the input layer is the same as the number of impedance features extracted. There are multiple hidden layers, and each hidden layer contains a certain number of neurons. The number of neurons in the output layer can be designed according to the specific requirements of the task. For a binary classification task (such as whether it is a tumor), the output layer can have 1 neuron and use the sigmoid activation function; for a multi-classification task (such as tumor type classification), the number of neurons in the output layer is equal to the number of classes and uses the softmax activation function.

[0041] Use the processed case data as the training set, and use the stochastic gradient descent method to preliminarily train the neural network structure; Obtain the number of training rounds of the neural network structure, and when the number of training rounds reaches the preset threshold, use batch gradient descent to optimize and train the neural network structure until the training ends.

[0042] Specifically, by combining the advantages of stochastic gradient descent and batch gradient descent, calculating the gradient for a small batch of data can accelerate training and avoid overfitting. S-BGD (Stochastic Batch Gradient Descent) can improve the training speed and stability through small batch updates. When the batch size during training is large or the memory is limited, the gradient accumulation strategy can alleviate the memory limitation problem. By accumulating gradients after multiple small batch trainings and then performing a single weight update, large batch training can be achieved without increasing memory usage.

[0043] Update the neural network parameters through the backpropagation algorithm to obtain the trained neural network model, and use it as the AI tumor recognition model.

[0044] Use the impedance feature parameters as the input, and use the AI tumor recognition model to infer and output the corresponding breast tissue classification result; Judge the abnormality of breast tissue and analyze the abnormal type and degree according to the result output by the AI tumor recognition model; It should be noted that based on the output result of the AI model, judge the abnormality of breast tissue. The abnormal conditions include: analysis of abnormal type and degree; For example, judge whether the breast tissue is a tumor according to the output classification result of the model, and the type of tumor (such as benign or malignant).

[0045] Through the probability output provided by the AI model, the severity of the tumor can be further analyzed. For example, the output may include information such as the malignancy probability of the tumor, the size and location of the tumor.

[0046] Analyze the combination of impedance feature parameters, and restore the complete breast detection result according to the judgment and analysis results.

[0047] Specifically, the impedance characteristic parameters are combined and analyzed with the diagnostic results output by the model. The AI model not only provides classification results, but also provides some key diagnostic parameters, such as tumor size, shape, etc. Through this information, the complete results of breast detection can be restored to help doctors make subsequent diagnostic and treatment decisions.

[0048] A visualization display module is used to display the original breast detection result information on the software interface and present it to the user in the form of intuitive charts (such as impedance distribution maps, tumor location schematic diagrams, etc.), images (such as breast images superimposed with detection results), or text. A report generation module is used to generate standardized or personalized breast detection reports according to the user's selection.

[0049] It should be noted that the report content includes the detection time, details of various detection results, and further medical measures recommended for the detection results, etc., and the report can be exported to common document formats such as PDF and Word, which is convenient for users to save, print, and carry for doctors to view.

[0050] As Figure 3 shown, according to another embodiment of the present invention, a detection method for early breast cancer detection based on AI analysis is provided, including the following steps: S1. Control the signal emission module on the bra main body through the mobile phone to emit low-voltage high-frequency alternating current and output it to the EIT sensor. S2. Collect breast tissue impedance information through the EIT sensor. S3. Preprocess the breast tissue impedance information, and extract impedance characteristic parameters related to the physiological state of the breast tissue from the preprocessed digital signal. S4. Encode and modulate the impedance characteristic parameters into wireless signals and transmit them to an external receiving device, and at the same time manage the power supply and power monitoring. S5. Develop a signal emission plan, receive the impedance characteristic parameters, and identify abnormal conditions of the breast tissue based on a deep learning algorithm, and generate a visualization report according to the identification results.

[0051] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0052] The specific embodiments described above further elaborate on the objective, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only for the specific embodiments of the present invention and is not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A bra for early breast cancer detection based on AI analysis, characterized in that, Including: A signal transmission module, which is used to control, through a mobile phone, the signal transmission module located on the bra main body to emit low-voltage high-frequency alternating current and output it to the EIT sensor; The bra main body, which is used to collect the impedance information of breast tissue through the EIT sensor; A signal analysis module, which is used to preprocess the impedance information of breast tissue and extract impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed digital signal; A communication transmission device, which is used to encode and modulate the impedance characteristic parameters into a wireless signal and transmit it to the mobile phone, and manage the power supply and power monitoring at the same time; AI data analysis software, which is used to formulate a signal transmission plan, receive the impedance characteristic parameters, identify abnormal conditions of breast tissue based on a deep learning algorithm, and generate a visual report according to the identification result.

2. The bra for early breast cancer detection based on AI analysis according to claim 1, characterized in that, The bra main body is made of a flexible and breathable medical-grade elastic fabric, and an installation cavity is provided inside the bra main body. A number of EIT sensor arrays are arranged inside the installation cavity. Shoulder straps are also provided on both sides of the bra main body, and a power supply module is arranged on the outer side of one of the shoulder straps. One side of the power supply module is the signal transmission module, and the signal transmission module is connected to one of the shoulder straps; The signal transmission module is used to receive the control instruction from the mobile phone and transmit a low-voltage high-frequency alternating current signal.

3. The bra for early breast cancer detection based on AI analysis according to claim 2, characterized in that, The signal analysis module includes: A signal preprocessing unit, which is electrically connected to the EIT sensor array, receives and preprocesses the impedance information. The preprocessing includes filtering, amplification, and analog-to-digital conversion; A feature extraction unit, which is used to extract impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information.

4. A bra for early breast cancer detection based on AI analysis according to claim 3, characterized in that, The preprocessing of the impedance information includes: Filtering the impedance information based on wavelet transform and Wiener filtering method; Amplifying the impedance information after filtering using a low-noise amplifier; Converting the impedance information after amplification into a digital signal using a mode converter.

5. A bra for early breast cancer detection based on AI analysis according to claim 4, characterized in that, The extraction of impedance characteristic parameters related to the physiological state of breast tissue from the preprocessed impedance information includes: Performing a first-level resolution discrete Fourier transform on the preprocessed impedance signal, and calculating the center frequency and determining the frequency range where the power drops to half of the peak value according to the discrete Fourier transform result; Based on the center frequency and the frequency range where the power drops to half of the peak value, optimizing the discrete Fourier transform parameters and performing a second-level resolution discrete Fourier transform on the impedance signal; Extracting impedance characteristics related to the physiological state of breast tissue from the high-resolution DFT result.

6. The bra for early breast cancer detection based on AI analysis according to claim 1, wherein The AI data analysis software includes: A signal transmission control module, which is used to formulate a signal transmission plan and transmit it to the signal transmission module on the bra main body; A data transmission and reception module, which is used to transmit a signal control instruction through the communication transmission device, receive the transmitted impedance characteristic parameters, and perform temporary storage; An AI tumor recognition module for inputting impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judging breast tissue abnormality, analyzing the type and degree of abnormality according to the output result of the AI tumor recognition model, and parsing the impedance characteristic parameters to restore the original breast detection result information; A visualization display module for displaying the original breast detection result information on the software interface and presenting it to the user in an intuitive chart, image or text form; A report generation module for generating a standardized or personalized breast detection report according to the user's selection.

7. A bra for early breast cancer detection based on AI analysis according to claim 6, characterized in that, The process of inputting the impedance characteristic parameters into a pre-trained AI tumor recognition model for calculation, judging breast tissue abnormality, analyzing the type and degree of abnormality according to the output result of the AI tumor recognition model, and parsing the impedance characteristic parameters to restore the original breast detection result information includes: Collecting breast case data, constructing and training a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model; Taking the impedance characteristic parameters as input, using the AI tumor recognition model to infer and output the corresponding breast tissue classification result; Judging breast tissue abnormality and analyzing the type and degree of abnormality according to the output result of the AI tumor recognition model; Parsing the combined impedance characteristic parameters, and restoring the complete breast detection result according to the judgment and analysis results.

8. A bra for early breast cancer detection based on AI analysis according to claim 7, characterized in that, The process of collecting breast case data, constructing and training a neural network model based on a deep learning algorithm to obtain an AI tumor recognition model includes: Collecting breast case data and performing data processing. The breast case data includes impedance signals and case labels. The data processing includes: standardizing, denoising and feature extraction of the impedance signals; Constructing a neural network structure, using the processed case data, and iteratively training the neural network structure using the S-BGD and gradient accumulation strategies; Updating the neural network parameters through the backpropagation algorithm to obtain a trained neural network model, and using it as the AI tumor recognition model.

9. The bra for early breast cancer detection based on AI analysis according to claim 8, characterized in that, The process of constructing a neural network structure, using the processed case data, and iteratively training the neural network structure using the S-BGD and gradient accumulation strategies includes: Constructing the input layer, hidden layer and output layer of the neural network structure; Taking the processed case data as the training set and preliminarily training the neural network structure using the stochastic gradient descent method; Obtaining the training rounds of the neural network structure, and when the training rounds reach a preset threshold, optimizing and training the neural network structure using batch gradient descent until the training ends.

10. A detection method for early breast cancer detection based on AI analysis, which operates using a bra for early breast cancer detection based on AI analysis as described in any one of claims 1-9, characterized in that, Including the following steps: S1. Controlling the signal emission module on the bra main body through the mobile phone to emit a low-voltage high-frequency alternating current and output it to the EIT sensor; S2. Collecting breast tissue impedance information through the EIT sensor; S3. Preprocessing the breast tissue impedance information, and extracting impedance characteristic parameters related to the physiological state of the breast tissue from the preprocessed digital signal; S4. Encoding and modulating the impedance characteristic parameters into a wireless signal and transmitting it to the mobile phone, while managing the power supply and power monitoring. S5. Develop a signal emission plan, receive impedance characteristic parameters, identify abnormal conditions of breast tissue based on a deep learning algorithm, and generate a visualization report according to the identification results.

Images