Male premature ejaculation detection device and detection method

Abstract

The invention discloses a male premature ejaculation detection device and method, and relates to the field of human health detection equipment.The male premature ejaculation detection device comprises a physiological signal collection module and an intelligent evaluation terminal, the physiological signal collection module monitors the penis hardness and erection duration through a flexible lantern ring, and a detection module collects the movement track and the pressure change of a silica gel lantern ring; the detection head assembly measures a penis head biosensory threshold. A data processing module of the intelligent evaluation terminal integrates a data analysis unit, a data encryption unit and a self-adaptive calibration unit, the data analysis unit adopts an artificial neural network to realize premature ejaculation state evaluation classification, the data encryption unit encrypts and stores user data, and the self-adaptive calibration unit can adaptively optimize the sensitivity of a sensor. The premature ejaculation level is judged through the auxiliary evaluation module, detection data are visually displayed, and an evaluation report is generated. According to the method, multi-dimensional physiological parameter fusion collection and intelligent analysis can be realized, the evaluation accuracy is improved, the privacy security of the user is guaranteed, and the method has remarkable clinical application value.

Description

Technical Field

[0001] This application relates to the field of human health testing equipment technology, specifically to a male premature ejaculation testing device and testing method. Background Technology

[0002] Premature ejaculation (PE) is one of the most common male sexual dysfunctions in clinical practice, with an estimated global prevalence of 20-30%, seriously affecting patients' physical and mental health and their relationships. Currently, the clinical diagnosis of PE mainly relies on the definition from the International Society for Sexual Medicine (ISSM): always or almost always ejaculating within approximately one minute of penetration, completely or almost completely losing the ability to delay ejaculation, resulting in negative personal consequences (such as distress, irritability, frustration, and / or avoidance of sexual activity). The core of this definition lies in three dimensions: "short duration," "feeling of loss of control," and "negative impact." However, in clinical practice, relying solely on patients' verbal reports and questionnaires (such as the PEDT scale) to assess these three dimensions presents significant objectivity challenges. Furthermore, this subjective diagnostic method has obvious limitations: patients' descriptions may be inaccurate due to psychological factors, partner relationships, cognitive biases, etc., lacking objective and quantifiable diagnostic evidence.

[0003] Therefore, clinical diagnosis has long faced the dilemma of "having standards, but being difficult to measure," and there is an urgent need for a testing method that can provide objective and quantitative physiological evidence to assist doctors in making more accurate diagnoses and classifications.

[0004] Some attempts have been made to measure objective physiological parameters in existing technologies, but all of them have obvious limitations: The intravaginal ejaculation latency (IELT) stopwatch method provides an objective time, but it is inconvenient to operate in a clinical setting, is easily affected by various factors, and can only reflect the single dimension of "time". It cannot reveal other key physiological changes that occur during ejaculation, such as changes in erectile rigidity and the state of pelvic floor muscle activity.

[0005] Biothesiometry: This method measures the sensitivity of the glans penis through vibration or electrical stimulation, aiming to explore the potential cause of "sensory nerve hypersensitivity." However, numerous studies have shown that not all PE patients have sensory nerve hypersensitivity; it is only one of many potential factors. Therefore, relying solely on this parameter cannot comprehensively assess ejaculatory control ability and can easily lead to misdiagnosis due to oversimplification.

[0006] Digital Vibratory Sensory Threshold (DVST) test: DVST only assesses local sensory input to the penis and fails to cover key aspects such as central integration during ejaculation, execution of pelvic floor muscles, and erectile status. It cannot reveal a patient's muscle control, neural reflex patterns, or overall sensitivity of the ejaculatory reflex arc during sexual activity.

[0007] Surface electromyography (sEMG) of the pelvic floor muscles: This technique records the electrical activity of pelvic floor muscles (such as the bulbospongiosus muscle) related to ejaculation, reflecting the "muscle execution" process of ejaculation. However, in current applications, it is mostly used for biofeedback therapy and rarely integrated into the diagnostic process. Furthermore, it is not analyzed synchronously with other key parameters such as erectile status and nerve sensitivity, resulting in information silos.

[0008] Therefore, there is a need for a detection device that can integrate multiple key physiological parameters and provide objective, quantitative, and comprehensive diagnostic evidence to overcome the limitations of existing technologies that rely on subjective descriptions, have low accuracy, and detect only a single parameter. Summary of the Invention

[0009] In view of this, this application provides a device and method for detecting premature ejaculation in men, in order to solve the technical problems of the limitations of the prior art that rely on subjective description and single parameter detection.

[0010] To achieve the above objectives, the present invention provides the following technical solution: This invention mainly consists of a male premature ejaculation detection device and detection method. It can non-invasively, objectively, and quantitatively collect and analyze multiple key physiological signals related to male premature ejaculation, and perform comprehensive evaluation through an algorithm model, providing reliable quantitative evidence for clinical diagnosis. The male premature ejaculation detection device includes: a physiological signal acquisition module and an intelligent evaluation terminal. The intelligent evaluation terminal includes a data processing module and an auxiliary evaluation module.

[0011] The physiological signal acquisition module includes a flexible ring device, a detection module, and a detection head assembly. The flexible ring device is used to measure penile rigidity, erection duration, and electromyographic signals of the pelvic floor muscles in real time. It is a wearable, flexible ring-shaped sensing device that integrates miniature pressure and strain sensors. It can be worn snugly at the base of the penis to detect radial pressure reflecting rigidity and circumferential deformation to assist in verifying erection. The detection module is connected to the flexible ring device and is used to acquire data on penile movement trajectory and pressure changes within the silicone ring. The detection head assembly is used to quantitatively measure the biosensory threshold of the glans penis. This assembly includes a standardized miniature vibrator and a microcurrent stimulator. The microcurrent stimulator uses a constant current within a safe microcurrent range.

[0012] The data processing module includes a data analysis unit, a data encryption unit, and an adaptive calibration unit. The data analysis unit is electrically connected to the physiological signal acquisition module. It receives the acquired physiological signal data, filters and extracts features from it, and uses an artificial neural network-based assessment and classification model to assess and classify premature ejaculation based on the extracted feature parameters, outputting the analysis results. The data encryption unit encrypts and stores user data. The adaptive calibration unit optimizes the sensitivity of each sensor based on historical data using machine learning algorithms. The auxiliary assessment module is used to automatically determine the level of premature ejaculation based on preset thresholds and analysis results, and to visualize the test data and output a test analysis report.

[0013] As a further aspect of the present invention: the flexible collar device includes a silicone collar, a micro pressure sensor array, a strain sensor, a surface electromyography (SEMG) electrode, and a flexible circuit board. The micro pressure sensor array, strain sensor, and SEMG electrode are integrated on the surface of the silicone collar. The flexible circuit board is embedded inside the silicone collar and includes a temperature compensation module and a timing module. The temperature compensation module is used to dynamically adjust the zero-point calibration parameters of the pressure sensor according to the ambient temperature, and the timing module is used to record time information. The SEMG electrode is integrated on the inner surface of the ring sensor and contacts the skin. It is used to collect electrical activity signals of muscles such as the bulbospongiosus muscle and ischiocavernosus muscle under sexual stimulation. Signal interference is eliminated by a preamplifier and a bandpass filter set on the flexible circuit board.

[0014] As a further aspect of the present invention: the detection module includes a housing and a substrate disposed within the housing. The substrate includes a triaxial gravity sensor, a pressure sensor, a data communication module, and a power management chip. The triaxial gravity sensor is used to detect the frequency, amplitude, and direction changes of the penis's swaying motion. The pressure sensor is used to monitor pressure changes within the silicone collar, forming a redundancy check with the miniature pressure sensor array. The data communication module is used to upload the detected pressure, motion trajectory, and pressure data to the main control module. The power management chip is used to dynamically manage the status of the lithium battery, which powers the detection module and the flexible collar device. The data communication module includes an MCU module and a dual-mode communication module. The MCU uses an ARM Cortex-M4 core, the gravity sensor communicates with the MCU via an I²C bus, and the pressure sensor transmits data via an SPI bus. The dual-mode communication module adopts a dual-mode communication architecture, supporting both low-power Bluetooth 5.2 and high-speed Wi-Fi protocols. During the data acquisition phase, the dual-mode communication module can establish a low-latency connection with the user's mobile device via Bluetooth 5.2 to transmit pressure, motion, and air pressure data. During the data synchronization phase, the data is uploaded to the server via Wi-Fi, supporting remote diagnostics and data sharing. The module also has a built-in AES-256 encryption chip to ensure the confidentiality and integrity of data transmission.

[0015] As a further embodiment of the present invention: the detection head assembly includes a main control module, a handheld feedback button, a replaceable medical silicone sleeve, and a detection head encapsulated in the medical silicone sleeve. The inner wall of the detection head is provided with multiple detection ports along the axial direction. Each detection port has a built-in micro-vibrator and a micro-current stimulator. The handheld feedback button, the micro-vibrator, and the micro-current stimulator are connected to the main control module through connecting wires. The main control module is used to control the vibration frequency and current intensity. The main control module is connected to the data processing module.

[0016] As a further aspect of the present invention: the data analysis unit includes a feature extraction module and an analysis module; The feature extraction module is used to process the acquired physiological signals and extract multivariate feature parameters, which include time dimension feature parameters, intensity dimension feature parameters and control dimension feature parameters. The analysis module is connected to the feature extraction module. The analysis module is used to use an evaluation classification model based on artificial neural networks to output the classification probability of premature ejaculation based on the input multivariate feature parameters, and transmit the classification probability value to the auxiliary evaluation module. The auxiliary evaluation module compares the preset threshold with the classification probability value to determine the level of premature ejaculation.

[0017] A detection method based on the above-mentioned male premature ejaculation detection device includes the following steps: Step S1: The user wears the physiological signal acquisition module. After the device is powered on, the main control module executes the sensor calibration program to adaptively calibrate each sensor. In the absence of sexual stimulation, the detection head assembly is used to measure the user's penile head biosensory threshold. Step S2: Activate the standard audiovisual stimulation unit to present uniform sexual stimulation content to the user. Under sexual stimulation, simultaneously activate the flexible ring device and detection module to continuously collect data on penile rigidity and circumference changes, pelvic floor muscle electromyography data, and movement trajectory data, and start timing. Step S3: Determine the erection onset time T1 based on penile rigidity data, determine the ejaculation onset time T2 based on pelvic floor muscle electromyography data and / or pressure change data, and calculate the ejaculation latency IELT = T2 - T1; Step S4: Process the collected physiological signals and extract time-dimensional, intensity-dimensional, and control-dimensional feature parameters. Time-dimensional feature parameters include ejaculation latency (IELT), duration of penile maximal rigidity, and duration of rhythmic contractions of the pelvic floor muscles. Intensity-dimensional feature parameters include the maximum penile rigidity value, the average amplitude and integral electromyographic value of pelvic floor muscle signals before and during ejaculation, and the complexity of the movement trajectory. Control-dimensional feature parameters include pelvic floor muscle electromyographic activity characteristics and features of the penile glans biosensory threshold. The extracted multi-dimensional feature parameters are input into an assessment and classification model based on an artificial neural network. The model outputs the classification probability of premature ejaculation based on the input multi-dimensional feature parameters and transmits the classification probability value to the auxiliary assessment module. The auxiliary assessment module compares the preset threshold with the classification probability value to determine the level of premature ejaculation.

[0018] As a further aspect of the present invention, the electromyographic activity characteristics of the pelvic floor muscles include the stability index of the electromyographic signal before ejaculation, and the frequency and amplitude of involuntary spasmodic contractions that occur during sexual stimulation.

[0019] As a further aspect of the present invention: measuring the biosensory threshold of a user's penile head using a detection head assembly includes: the main control module controlling a micro vibrator and a microcurrent stimulator to sequentially apply a series of vibration stimuli and microcurrent stimuli with increasing intensity; The system acquires the feedback signal when the user presses the handheld feedback button when they first feel the stimulus, and records the minimum stimulus intensity when the user feels the stimulus, which are used as the vibration sensation threshold and the current sensation threshold, respectively.

[0020] As a further aspect of the present invention: determining the erection onset time T1 based on penile rigidity data, and determining the ejaculation occurrence time T2 based on pelvic floor muscle electromyography data and / or pressure change data includes: Real-time monitoring of penile rigidity data; when the rigidity reaches the preset erection pressure threshold, record the time point T1. Continuously collect pressure change data and pelvic floor muscle electromyography (EMG) data, and analyze the pelvic floor muscle EMG signals in real time. When the corresponding rhythmic strong contraction of the pelvic floor muscle group EMG signal and pressure change signal are detected, the ejaculation event is determined to have occurred, and the time point T2 of the ejaculation event is recorded.

[0021] As a further aspect of the present invention: the evaluation classification model based on artificial neural networks includes: an input layer, the number of which is equal to the dimension of multiple feature parameters; At least one hidden layer, which uses ReLU as the activation function; An output layer with the number of neurons equal to the number of classification categories, and using the Softmax activation function to output the probability of belonging to each category; Artificial neural networks use classification cross-entropy as the loss function and are trained using the Adam optimizer.

[0022] As can be seen from the above technical solution, the advantages of the present invention are: This application utilizes multi-sensor fusion information acquisition to obtain multiple quantifiable physiological parameters, providing objective evidence for diagnosis. It organically combines three key parameters—penile state characteristics, glans sensitivity, and pelvic floor muscle electromyographic activity—to comprehensively assess the physiological basis of premature ejaculation from two dimensions: nerve sensitivity and muscle control ability. Furthermore, by employing a machine learning model to comprehensively analyze these multiple parameters, it can more accurately distinguish between physiological and psychologically-driven premature ejaculation, and help identify different subtypes, providing direction for personalized treatment. In addition, the use of standardized stimulation sources and detection procedures reduces detection errors caused by differences in environment and stimulation content, resulting in better repeatability and comparability of the results. Attached Figure Description

[0023] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0024] Figure 1 This is a schematic diagram of the overall structure of this application.

[0025] Figure 2 This is a structural schematic diagram of the flexible collar device and detection module of this application.

[0026] Figure 3 This is a schematic diagram of the detection head assembly of this application.

[0027] Figure 4 This is a schematic diagram illustrating the specific steps of the male premature ejaculation detection method described in this application.

[0028] Explanation of reference numerals in the attached drawings: 1. Flexible collar device; 2. Detection module. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and their descriptions are used to explain this application, but are not intended to limit it.

[0030] refer to Figures 1 to 4 This embodiment provides a male premature ejaculation detection device and method capable of non-invasively, objectively, and quantitatively collecting and analyzing multiple key physiological signals related to premature ejaculation, and comprehensively evaluating them through an algorithm model to provide reliable quantitative evidence for clinical diagnosis. The male premature ejaculation detection device includes: a physiological signal acquisition module and an intelligent assessment terminal. The intelligent assessment terminal includes a data processing module and an auxiliary assessment module. Figure 2 As shown, specifically, the physiological signal acquisition module includes a flexible collar device 1, a detection module 2, and a detection head assembly. The flexible collar device 1 is used to measure the hardness and duration of penile erection and the electromyographic signals of the pelvic floor muscle group in real time. The detection module 2 is connected to the flexible collar device 1 and is used to collect data on penile movement trajectory and pressure changes within the silicone collar. The detection head assembly is used to measure the biosensory threshold of the penile head region.

[0031] In this embodiment, the flexible collar device 1 is a wearable, flexible ring-shaped sensing device, including a silicone collar, a micro pressure sensor array, a strain sensor, surface electromyography (EMG) electrodes, and a flexible circuit board. The silicone collar is made of medical-grade silicone, with two Ag / AgCl silver chloride surface EMG electrodes positioned on its inner wall where it contacts the skin for pelvic floor EMG signal acquisition. The micro pressure sensor array and strain sensor are integrated on the surface of the silicone collar, and the flexible circuit board is embedded inside the silicone collar. The pressure sensor array uses a MEMS pressure sensor, and the strain sensor is a flexible strain gauge. The flexible circuit board includes a temperature compensation module and a timing module. The temperature compensation module dynamically adjusts the zero-point calibration parameters of the pressure sensor according to the ambient temperature, and the timing module records time information. The micro pressure sensor array detects radial pressure reflecting rigidity, the strain sensor detects circumferential deformation information to assist in verifying erectile function, and the surface EMG electrodes contact the skin to collect electrical activity signals from muscles such as the bulbospongiosus and ischiocavernosus muscles under sexual stimulation. Signal interference is eliminated by a preamplifier and a bandpass filter mounted on the flexible circuit board.

[0032] The detection module 2 is mounted on the flexible collar device 1. The detection module 2 includes a housing and a base plate disposed within the housing, made of medical-grade ABS plastic. The base plate includes a triaxial gravity sensor, a pressure sensor, a data communication module, and a power management chip. The triaxial gravity sensor is used to detect changes in the frequency, amplitude, and direction of penile oscillation. The pressure sensor is used to monitor pressure changes within the silicone collar and identify pressure abrupt changes. This sensor can form a redundancy check with the pressure sensor array to improve data reliability. The data communication module is used to upload the detected pressure, motion trajectory, and pressure data to the main control module. The power management chip is used to dynamically manage the status of the power supply lithium battery.

[0033] In this embodiment, the data communication module includes an MCU module and a dual-mode communication module. The MCU uses an ARM Cortex-M4 core, the gravity sensor communicates with the MCU via the I²C bus, and the barometric pressure sensor transmits data via the SPI bus. The dual-mode communication module adopts a dual-mode communication architecture, supporting both low-power Bluetooth 5.2 and high-speed Wi-Fi protocols. During the data acquisition phase, the dual-mode communication module can establish a low-latency connection with the user's mobile device via Bluetooth 5.2 to transmit pressure, motion, and barometric pressure data. During the data synchronization phase, data is uploaded to the server via Wi-Fi, supporting remote diagnostics and data sharing. The module also incorporates an AES-256 encryption chip to ensure the confidentiality and integrity of data transmission.

[0034] The detection head assembly is a standalone handheld component, comprising a main control module, a handheld feedback button, a disposable medical silicone sleeve, and a detection head encapsulated within the sleeve. The inner wall of the detection head has two axially aligned detection ports, each housing a standard-sized miniature vibrator and a microcurrent stimulator. The microcurrent stimulator utilizes a constant current source circuit, with a maximum output current limited to below 1mA to ensure safety. The handheld feedback button, miniature vibrator, and microcurrent stimulator are connected to the main control module via connecting cables. The main control module controls the vibration frequency and current intensity. It is also connected to a data processing module, and the microcontroller used in the main control module is an STMicroelectronics STM32L4 series microcontroller.

[0035] The data processing module includes a data analysis unit, a data encryption unit, and an adaptive calibration unit. The data analysis unit is electrically connected to the physiological signal acquisition module. It receives the acquired physiological signal data, filters and extracts features from it, and uses an artificial neural network-based assessment and classification model to classify and assess premature ejaculation based on the extracted feature parameters, outputting the analysis results. The data encryption unit encrypts and stores user data. The adaptive calibration unit optimizes the sensitivity of each sensor based on historical data using machine learning algorithms. The data analysis unit includes a feature extraction module and an analysis module. The feature extraction module processes the acquired physiological signals and extracts multivariate feature parameters, including time-dimension, intensity-dimension, and control-dimension features. The analysis module is connected to the feature extraction module and uses an artificial neural network-based assessment and classification model to output the classification probability of premature ejaculation based on the input multivariate feature parameters. This classification probability value is then transmitted to the auxiliary assessment module, which compares a preset threshold with the classification probability value to determine the level of premature ejaculation.

[0036] The auxiliary assessment module is used to automatically determine the level of premature ejaculation based on preset thresholds and analysis results, and to visualize the test data and output a test analysis report. The test analysis report includes objectively measured IELT values, penile rigidity curves, pelvic floor muscle electromyography, glans penis biosensory thresholds, comprehensive assessment conclusions and risk levels, and possible subtype indications, such as based on excessive sensitivity or poor pelvic floor muscle control.

[0037] This application also discloses a detection method based on the above-mentioned male premature ejaculation detection device, comprising the following steps: Step S1: Preparation and calibration stage. The user wears the physiological signal acquisition module in a private and comfortable environment. After the device is powered on, the main control module executes the sensor calibration program to adaptively calibrate each sensor. In the absence of sexual stimulation, the detection head assembly is used to measure the user's penile head biosensory threshold.

[0038] Measuring a user's penile head biosensory thresholds using a detection head assembly includes: Step S10: The main control module controls the micro vibrator and microcurrent stimulator to sequentially apply a series of vibration stimuli and microcurrent stimuli with increasing intensity; Step S11: Obtain the feedback signal when the user presses the handheld feedback button when they first feel the stimulus, and record the minimum stimulus intensity when the user feels the stimulus, which will be used as the vibration sensation threshold and the current sensation threshold, respectively.

[0039] Step S2: Standardized Stimulation and Real-time Data Acquisition Stage. The standard audiovisual stimulation unit is activated to present uniform sexual stimulation content to the user. Under sexual stimulation, the flexible ring device 1 and detection module 2 are activated simultaneously to continuously collect data on penile rigidity and circumference changes, pelvic floor muscle electromyography data, and movement trajectory data, and the timing is started. Step S3: Determine the erection onset time T1 based on penile rigidity data, determine the ejaculation onset time T2 based on pelvic floor muscle electromyography data and / or pressure change data, and calculate the ejaculation latency IELT = T2 - T1.

[0040] Specifically, determining the erection onset time T1 based on penile rigidity data and the ejaculation onset time T2 based on pelvic floor muscle electromyography data and / or pressure change data includes: Real-time monitoring of penile rigidity data; when the rigidity reaches the preset erection pressure threshold, record the time point T1. For example, the preset erection pressure threshold is equivalent to the pressure value of 80% full erection. Continuously collect pressure change data and pelvic floor muscle electromyography (EMG) data, and analyze the pelvic floor muscle EMG signals in real time. When the corresponding rhythmic strong contraction of the pelvic floor muscle group EMG signal and pressure change signal are detected, the ejaculation event is determined to have occurred, and the time point T2 of the ejaculation event is recorded.

[0041] Step S4: Data Processing and Feature Extraction Stage. The collected raw data is transmitted to the intelligent assessment terminal for feature extraction. The collected physiological signals are processed to extract time-dimensional, intensity-dimensional, and control-dimensional feature parameters. Time-dimensional feature parameters include ejaculation latency (IELT), duration of penile maximal rigidity, and duration of rhythmic contractions of the pelvic floor muscles. Intensity-dimensional feature parameters include the maximum penile rigidity value, the average amplitude and integral electromyographic value of pelvic floor muscle signals before and during ejaculation, and the complexity of the movement trajectory. Control-dimensional feature parameters include pelvic floor muscle electromyographic activity characteristics and features of the penile glans biosensory threshold. The extracted multiple feature parameters are input into an assessment classification model based on an artificial neural network. The model outputs the classification probability of premature ejaculation based on the input multivariate feature parameters and transmits the classification probability value to the auxiliary assessment module. The auxiliary assessment module compares the preset threshold with the classification probability value to determine the level of premature ejaculation.

[0042] The electromyographic characteristics of the pelvic floor muscles include stability indicators of preejaculate electromyographic signals and the frequency and amplitude of involuntary spasmodic contractions during sexual stimulation. These stability indicators include the root mean square or standard deviation of the signal in that frequency band.

[0043] The evaluation classification model based on artificial neural networks includes: an input layer with the number of neurons equal to the dimension of multiple feature parameters; at least one hidden layer using ReLU as the activation function; an output layer with the number of neurons equal to the number of classification categories and using the Softmax activation function to output the probability of belonging to each category; the artificial neural network uses classification cross-entropy as the loss function and is trained using the Adam optimizer.

[0044] In this embodiment, the evaluation classification model adopts a feedforward structure consisting of an input layer, two hidden layers, and an output layer, as detailed below: (1) Input layer, number of neurons: strictly corresponds to the dimension of the feature vector. Based on the 8 core features extracted above, the input layer is fixed to 8 neurons.

[0045] Input values: Each neuron receives a feature value standardized by Z-score. That is: Neuron_1: Standardized IELT value. Neuron_2: Standardized time to reach maximum hardness (T2) Neuron_3: Standardized maximum hardness value (H1) Neuron_4: Standardized pelvic floor muscle stability (E1) Neuron_5: Standardized frequency of spastic contractions (E2) Neuron_6: Standardized ejaculatory electromyographic intensity (E3) Neuron_7: Standardized vibration sensory threshold (S1) Neuron_8: Standardized electrosensory threshold (S2).

[0046] (2) Hidden layers: This invention uses two hidden layers to achieve a balance between model complexity and generalization ability. Number of neurons in the first hidden layer: 64.

[0047] The activation function is ReLU, mathematically expressed as f(x) = max(0, x). ReLU effectively alleviates the vanishing gradient problem, is computationally simple, and introduces the non-linearity required by the network. This layer receives an 8-dimensional vector from the input layer, performs a linear transformation using a weight matrix W^(1) (8x64 dimensions) and a bias vector b^(1) (64x1 dimensions), and then applies the ReLU activation function to the result. The output is a 64-dimensional vector.

[0048] Number of neurons in the second hidden layer: 32.

[0049] Activation function: ReLU, maintaining consistency while introducing non-linearity. This layer receives a 64-dimensional vector from the output of the first hidden layer, performs a linear transformation using a weight matrix W^(2) (64x32 dimension) and a bias vector b^(2) (32x1 dimension), and then applies the ReLU activation function. The output is a 32-dimensional vector.

[0050] After each hidden layer, a Dropout layer can be added as a regularizer to prevent overfitting.

[0051] (3) Output layer, number of neurons: For a three-classification task, there are 3 neurons, classifying premature ejaculation as healthy, lifelong premature ejaculation, and acquired premature ejaculation. Activation function: Softmax.

[0052] The three neurons in the output layer represent the predicted probabilities of belonging to the three categories: "healthy," "lifelong premature ejaculation," and "acquired premature ejaculation," respectively. For example, the output might be [0.05, 0.80, 0.15], indicating that the model judges the sample to have a 5% probability of being healthy, an 80% probability of having lifelong premature ejaculation, and a 15% probability of having acquired premature ejaculation. The final classification result is the category with the highest probability.

[0053] The implementation process of the detection method is as follows: One participant was tested.

[0054] In phase S100, the subject puts on the device and the system is calibrated.

[0055] During the S200 phase, the glans penis vibration sensory threshold was measured to be 2.5 μm, and the current sensory threshold was measured to be 0.3 mA.

[0056] Upon entering the S300 stage, a standardized video begins playing, and a timer starts. Approximately 120 seconds later (T1), the pressure sensor data indicates that the erection hardness has reached the erection threshold.

[0057] During the S400 phase, monitoring continued. Approximately 65 seconds after T1, i.e., the 185th second of the total time (T2), a characteristic high-frequency, high-amplitude rhythmic burst of electromyographic signals appeared, and the user simultaneously pressed the feedback button. The system determined that ejaculation had occurred, calculated the IELT to be 65 seconds, and stopped stimulation.

[0058] In the S500 stage, data is transmitted wirelessly to the intelligent assessment terminal. The intelligent assessment terminal extracts the following features: IELT=65s, maximum rigidity is 92%, pre-ejaculation pelvic floor electromyographic stability is poor, that is, the standard deviation is higher than the norm of healthy people, and the number of spastic contractions is 3 times / minute.

[0059] In the S600 stage, the above features are input into the trained artificial neural network model. The model comprehensively judges: short IELT, poor pelvic floor muscle control, and normal glans sensitivity. The output result is "Suspected acquired premature ejaculation, highly likely pelvic floor muscle neuromuscular control dysfunction," and a comprehensive score of "high risk" is given.

[0060] This invention enables objective, quantitative, and multi-dimensional diagnosis of premature ejaculation, overcoming the shortcomings of existing technologies that rely on subjective descriptions, and providing a precise basis for clinical classification and treatment.

[0061] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. In some cases, the actions or steps recorded in the specification and claims can be performed in a different order than that shown in the embodiments, and the desired result can still be achieved. In addition, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result; in some embodiments, multitasking and parallel processing are also feasible or advantageous.

[0062] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing the relevant hardware to implement them. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0063] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to the embodiments of this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A device for detecting premature ejaculation in men, characterized in that, include: A physiological signal acquisition module and an intelligent assessment terminal, wherein the intelligent assessment terminal includes a data processing module and an auxiliary assessment module; The physiological signal acquisition module includes a flexible collar device (1), a detection module (2), and a detection head assembly. The flexible collar device (1) is used to detect the hardness of the penis, the duration of erection, and the electromyographic signals of the pelvic floor muscle group. The detection module (2) is connected to the flexible collar device (1) and is used to collect data on penile movement trajectory and pressure changes within the silicone collar. The detection head assembly is used to measure the biosensory threshold of the penile head region. The data processing module includes a data analysis unit, a data encryption unit, and an adaptive calibration unit. The data analysis unit is electrically connected to the physiological signal acquisition module. The data analysis unit receives the acquired physiological signal data, filters and extracts features from the data, and uses an artificial neural network-based assessment and classification model to assess and classify premature ejaculation based on the extracted feature parameters, outputting the analysis results. The data encryption unit encrypts and stores user data. The adaptive calibration unit optimizes the sensitivity of each sensor based on historical data using machine learning algorithms. The auxiliary assessment module is used to automatically determine the level of premature ejaculation based on preset thresholds and analysis results, and to visualize the detection data and output a detection analysis report.

2. The male premature ejaculation detection device according to claim 1, characterized in that, The flexible collar device (1) includes a silicone collar, a micro pressure sensor array, a strain sensor, a surface electromyography electrode, and a flexible circuit board. The micro pressure sensor array, the strain sensor, and the surface electromyography electrode are integrated on the surface of the silicone collar. The flexible circuit board is embedded inside the silicone collar. The flexible circuit board includes a temperature compensation module and a timing module. The temperature compensation module is used to dynamically adjust the zero-point calibration parameters of the pressure sensor according to the ambient temperature. The timing module is used to record time information.

3. The male premature ejaculation detection device according to claim 1, characterized in that, The detection module (2) includes a housing and a base plate disposed within the housing. The base plate includes a triaxial gravity sensor, a pressure sensor, a data communication module, and a power management chip. The triaxial gravity sensor is used to detect the frequency, amplitude, and direction of penile swaying. The pressure sensor is used to monitor pressure changes within the silicone ring and forms a redundancy check with the miniature pressure sensor array. The data communication module is used to upload the detected pressure, motion trajectory, and pressure data to the main control module. The power management chip is used to dynamically manage the status of the power supply lithium battery.

4. The male premature ejaculation detection device according to claim 3, characterized in that, The detection head assembly includes a main control module, a handheld feedback button, a replaceable medical silicone sleeve, and a detection head encapsulated within the medical silicone sleeve. The inner wall of the detection head has multiple detection ports along the axial direction. Each detection port houses a miniature vibrator and a microcurrent stimulator. The handheld feedback button, the miniature vibrator, and the microcurrent stimulator are connected to the main control module via connecting wires. The main control module is used to control the vibration frequency and current intensity. The main control module is connected to the data processing module.

5. The male premature ejaculation detection device according to claim 1, characterized in that, The data analysis unit includes a feature extraction module and an analysis module; The feature extraction module is used to process the collected physiological signals and extract multivariate feature parameters, which include time dimension feature parameters, intensity dimension feature parameters and control dimension feature parameters. The analysis module is connected to the feature extraction module. The analysis module is used to use an evaluation classification model based on artificial neural networks to output the classification probability of premature ejaculation based on the input multivariate feature parameters, and transmit the classification probability value to the auxiliary evaluation module. The auxiliary evaluation module compares the preset threshold with the classification probability value to determine the level of premature ejaculation.

6. A detection method for male premature ejaculation detection device based on any one of claims 1 to 5, characterized in that, Includes the following steps: S1: The user wears the physiological signal acquisition module. After the device is powered on, the main control module executes the sensor calibration program to adaptively calibrate each sensor. In the absence of sexual stimulation, the detection head assembly is used to measure the user's penile head biosensory threshold. S2: Under sexual stimulation, the flexible ring device (1) and the detection module (2) are activated simultaneously to continuously collect penile rigidity data, pelvic floor muscle electromyography data, and movement trajectory data, and start timing; S3: Determine the erection onset time T1 based on the penile rigidity data, determine the ejaculation occurrence time T2 based on the pelvic floor muscle electromyography data and / or pressure change data, and calculate the ejaculation latency period IELT = T2 - T1. S4: Process the collected physiological signals and extract time-dimensional, intensity-dimensional, and control-dimensional feature parameters. The time-dimensional feature parameters include ejaculation latency (IELT), the duration of penile maximal rigidity, and the duration of rhythmic contractions of the pelvic floor muscles. The intensity-dimensional feature parameters include the maximum penile rigidity value, the average amplitude and integral electromyographic value of the pelvic floor muscles before and during ejaculation, and the complexity of the movement trajectory. The control-dimensional feature parameters include the electromyographic activity characteristics of the pelvic floor muscles and the feature parameters of the penile head biosensory threshold. The extracted multiple feature parameters are input into an evaluation and classification model based on an artificial neural network. Based on the input multivariate feature parameters, the model outputs the classification probability of premature ejaculation and transmits the classification probability value to the auxiliary evaluation module. The auxiliary evaluation module compares the preset threshold with the classification probability value to determine the level of premature ejaculation.

7. The method for detecting premature ejaculation in men according to claim 6, characterized in that, The electromyographic characteristics of the pelvic floor muscles include the stability of electromyographic signals before ejaculation, and the frequency and amplitude of involuntary spasmodic contractions that occur during sexual stimulation.

8. The method for detecting premature ejaculation in men according to claim 6, characterized in that, Measuring the user's penile head biosensory threshold using the detection head assembly includes: the main control module controlling the micro vibrator and the microcurrent stimulator to sequentially apply a series of vibration stimuli and microcurrent stimuli with increasing intensity; The system acquires the feedback signal when the user presses the handheld feedback button when they first feel the stimulus, and records the minimum stimulus intensity when the user feels the stimulus, which are used as the vibration sensation threshold and the current sensation threshold, respectively.

9. The method for detecting premature ejaculation in men according to claim 6, characterized in that, Determining the erection onset time point T1 based on the penile rigidity data, and determining the ejaculation onset time point T2 based on the pelvic floor muscle electromyography data and / or pressure change data, includes: Real-time monitoring of penile rigidity data; when the rigidity reaches the preset erection pressure threshold, the time point T1 is recorded. Continuously collect pressure change data and pelvic floor muscle electromyography (EMG) data, and analyze the pelvic floor muscle EMG signals in real time. When the corresponding rhythmic strong contraction of the pelvic floor muscle group EMG signal and pressure change signal are detected, the ejaculation event is determined to have occurred, and the time point T2 of the ejaculation event is recorded.

10. The method for detecting premature ejaculation in men according to claim 6, characterized in that, The evaluation classification model based on artificial neural networks includes: an input layer whose number of neurons is equal to the dimension of the multiple feature parameters; At least one hidden layer, wherein the hidden layer uses ReLU as the activation function; An output layer with the number of neurons equal to the number of classification categories, and using the Softmax activation function to output the probability of belonging to each category; The artificial neural network uses classification cross-entropy as the loss function and is trained using the Adam optimizer.

Images