Bladder urination tracking training system based on ultrasonic detection

Abstract

The invention discloses a bladder urination tracking training system based on ultrasonic detection, and belongs to the technical field of bladder urination training, and the system comprises an image collection module which is configured to collect a real-time image of the bladder of a user; the image processing module is configured to perform preprocessing, region segmentation and feature extraction on the real-time image of the bladder of the user; the bladder evaluation module is configured to construct a bladder function evaluation model to perform analysis and biological feature recognition on the bladder feature data of the user and determine a bladder function evaluation result of the user; and the urination training module is configured to perform urination tracking training on the bladder of the user. The problems that effective urination tracking training cannot be performed on the bladder of the user based on ultrasonic detection and the bladder control ability of the user cannot be effectively improved in the prior art, so that the bladder urination tracking training effect of the user is poor are solved. Effective urination tracking training can be performed on the bladder of the user based on ultrasonic detection, the bladder control ability of the user can be effectively improved, and the bladder urination tracking training effect of the user is good.

Description

Technical Field

[0001] The present invention relates to the technical field of bladder urination training, and specifically to a bladder urination tracking training system based on ultrasonic detection. Background Art

[0002] Bladder urination training is a behavioral therapy method aimed at helping users improve bladder control ability, reduce symptoms such as urinary incontinence, frequent urination, and urgency, and restore normal urination patterns. It helps users re-establish the storage and urination functions of the bladder through a systematic training plan and is applicable to patients with various bladder dysfunctions.

[0003] A Chinese patent with the publication number CN216294807U discloses a bladder urination training device, belonging to the technical field of medical devices. The urination training device includes a urination component and a control component. One end of the installation pipe is hermetically sleeved on the first branch pipe. An isolation diaphragm is fixedly installed inside the installation pipe. There is an oil storage chamber between the isolation diaphragm and the pressure sensor. The opposite sides of the first clamp block and the second clamp block are respectively provided with a first card slot and a second card slot. The middle part of the urination pipe is arranged between the first card slot and the second card slot. A sliding groove and an installation groove are opened inside the second clamp block. A telescopic member is fixedly installed inside the installation groove. A clamping plate is fixedly installed on the telescopic rod of the telescopic member. The clamping plate is slidably installed in the sliding groove. The telescopic member is electrically connected to the pressure sensor; it saves the use cost of the bladder urination training device and improves the cleanliness of the bladder urination training device. However, this patent has the following defects: The existing technology cannot effectively perform urination tracking training on the user's bladder based on ultrasonic detection, and cannot effectively improve the user's bladder control ability, resulting in poor effects of the user's bladder urination tracking training. Summary of the Invention

[0004] The purpose of the present invention is to provide a bladder urination tracking training system based on ultrasonic detection, which can effectively perform urination tracking training on the user's bladder based on ultrasonic detection, can effectively improve the user's bladder control ability, and has good effects on the user's bladder urination tracking training, solving the problem in the above-mentioned background art that the existing technology cannot effectively perform urination tracking training on the user's bladder based on ultrasonic detection, and cannot effectively improve the user's bladder control ability, resulting in poor effects of the user's bladder urination tracking training.

[0005] To achieve the above purpose, the present invention provides the following technical solutions: A bladder urination tracking training system based on ultrasonic detection, comprising: An image acquisition module configured to acquire real-time images of the user's bladder based on an ultrasonic probe; An image processing module configured to perform preprocessing, region segmentation, and feature extraction on the acquired real-time images of the user's bladder to determine the user's bladder feature data; Based on the standard key display feature information of each bladder feature index, verify the initial bladder features, and determine whether the clarity of the standard key display feature information in the initial bladder features is greater than the preset clarity, and whether the proportion of the standard key display feature information in the initial bladder features is greater than the preset proportion, so as to determine the user's bladder feature data; A bladder evaluation module, configured to construct a bladder function evaluation model to analyze and biometrically identify the user's bladder feature data, and determine the user's bladder function evaluation result; A urination training module, configured to formulate a user's bladder urination training plan according to the user's bladder function evaluation result, and conduct urination tracking training on the user's bladder.

[0006] Preferably, collect the user's real-time bladder image, including: Place the ultrasound probe of a medical ultrasound instrument on the lower abdomen of the user, and use the ultrasound probe to perform real-time detection and collection of the fullness degree, shape and structure of the user's bladder, so as to obtain the user's real-time bladder image based on ultrasound detection.

[0007] Preferably, preprocess the collected user's real-time bladder image, including: Denoise the user's real-time bladder image based on the Gaussian filtering algorithm, remove the speckle noise in the user's real-time bladder image, retain the user's bladder boundary information in the user's real-time bladder image, and reduce the noise interference in the user's real-time bladder image; Enhance the contrast of the user's real-time bladder image based on histogram equalization, highlight the user's bladder boundary in the user's real-time bladder image, and make the user's bladder area in the user's real-time bladder image clear; Standardize the user's real-time bladder image. Among them, perform gray normalization on the user's real-time bladder image, scale the pixel values in the user's real-time bladder image to a fixed range, and adjust the size of the user's real-time bladder image to unify the resolution, so as to eliminate the dimension difference in the user's real-time bladder image.

[0008] Preferably, perform region segmentation and feature extraction on the collected user's real-time bladder image, including: Based on the threshold segmentation method, set a threshold according to the gray value of the user's real-time bladder image, and perform region segmentation on the user's real-time bladder image according to the threshold to separate the bladder region; Extract features from the segmented bladder region, extract the features useful for bladder urination tracking training from the segmented bladder region, and determine the user's bladder feature data, including bladder volume feature, bladder wall thickness, residual urine volume, bladder shape feature, bladder wall movement feature, bladder internal echo feature and bladder pressure feature.

[0009] Preferably, extract the features useful for bladder urination tracking training from the segmented bladder region, including: Extract features of the segmented bladder region from texture and shape to obtain texture features and shape features; Determine key point information from preset standard bladder feature information, and extract features of the bladder region based on the key point information to obtain key point features; Performing feature clustering on the texture features, shape features and key point features respectively, and obtaining target texture features, target shape features and target key point features based on the feature clustering results; Based on the bladder feature index, the target texture feature, the target shape feature and the target key point feature are subjected to targeted feature fusion to obtain an initial bladder feature corresponding to each bladder feature index; Based on the standard key display feature information of each bladder feature indicator, the initial bladder feature is verified to determine whether the clarity of the standard key display feature information in the initial bladder feature is greater than the preset clarity, and whether the proportion of the standard key display feature information in the initial bladder feature is greater than the preset proportion; If yes, determining that the initial bladder feature verification is passed, and using the initial bladder feature as user bladder feature data; Otherwise, when the clarity of the standard key display feature information in the initial bladder feature is not greater than the preset clarity, feature enhancement processing is performed on the standard key display feature information in the initial bladder feature, and when the proportion of the standard key display feature information in the initial bladder feature is not greater than the preset proportion, other information in the initial bladder feature is eliminated to obtain the target bladder feature, and the target bladder feature is used as the user bladder feature data.

[0010] Preferably, constructing a bladder function assessment model comprises: According to the bladder urination tracking training requirements, bladder history data is collected and divided into a training set and a test set; Based on deep learning technology, a training set is used to train the deep learning model, so that the deep learning model can autonomously learn bladder function assessment behavior from the training set, and determine a bladder function assessment model based on deep learning; Based on the cross-validation method, the test set was used to test the performance of the deep learning-based bladder function assessment model, and to evaluate whether the deep learning-based bladder function assessment model can achieve the expected effect of assessing bladder function; According to the model test evaluation results, the parameters of the deep learning-based bladder function evaluation model are adjusted, and the deep learning-based bladder function evaluation model after parameter adjustment is continuously optimized until the deep learning-based bladder function evaluation model can achieve the expected effect of evaluating bladder function and determine the optimal bladder function evaluation model.

[0011] Preferably, the analysis and biometric recognition of the user's bladder characteristic data include: Deploy the optimal bladder function evaluation model in the actual bladder function evaluation environment; Input the user's bladder characteristic data into the bladder function evaluation model, analyze and perform biometric recognition on the user's bladder characteristic data according to the bladder function evaluation model, and evaluate the user's bladder function to determine the user's bladder function evaluation result, including the urine storage and urination functions of the user's bladder.

[0012] Preferably, the urination tracking training for the user's bladder includes: Formulate a urination training plan for the user's bladder according to the user's bladder function evaluation result; Conduct urination training on the user's bladder according to the formulated urination training plan for the user's bladder; Among them, the process of conducting urination training on the user's bladder is monitored in real time, the user's feedback is recorded, and when abnormal behavior is found, timely warnings are given, and the effect of the user's bladder urination training is tracked in real time, the effect of the user's bladder urination training plan is evaluated, and the user's bladder urination training plan is adjusted according to the evaluation effect and the user's feedback to optimize the user's bladder urination training plan.

[0013] Preferably, conducting urination training on the user's bladder according to the formulated urination training plan for the user's bladder includes: Formulate a scientific drinking plan for the user according to the user's bladder function evaluation result, including: the total daily water intake, the number of drinking times, the drinking time, and the single water intake; Guide the user to perform contraction and relaxation exercises of the pelvic floor muscles, each time lasting for 3 - 5 seconds, 3 - 5 groups are performed every day, and 10 - 15 times in each group, and the contraction of the user's pelvic floor muscles is monitored in real time, and the pelvic floor muscle exercise situation is adjusted in time according to the monitoring situation; Guide the user to urinate according to a fixed schedule, gradually extend the urination interval time, and stimulate the urination reflex by the sound of running water and the method of applying hot compress to the lower abdomen to enable the user to establish the awareness of autonomous urination.

[0014] Preferably, adjust the parameters of the bladder function evaluation model based on deep learning according to the model test evaluation result, including: Combined with the model test evaluation result, intelligently generate a grid including multiple parameter combinations of learning rate, number of layers, and number of neurons, and obtain the evaluation index under each model parameter combination based on grid search; Based on the evaluation index, calculate the comprehensive performance evaluation value of each model parameter combination; Based on the comprehensive performance evaluation value, combined with the model complexity under the model parameter combination, calculate the weighted weight of each model parameter combination; Select a target model parameter combination with a weighted weight greater than a preset weight, and perform weighted averaging on the prediction results under the target model parameter combination based on the weighted weight to obtain a final model prediction result.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention performs real-time detection and acquisition of the filling degree, shape, and structure of the user's bladder through an ultrasonic probe, obtains a real-time image of the user's bladder based on ultrasonic detection, preprocesses, regionally segments, and extracts features from the collected real-time image of the user's bladder to determine the user's bladder feature data, constructs a bladder function evaluation model according to the bladder urination tracking training requirements, analyzes and biometrically identifies the user's bladder feature data according to the bladder function evaluation model, evaluates the user's bladder function to determine the user's bladder function evaluation result, formulates a user's bladder urination training plan according to the user's bladder function evaluation result, and performs urination tracking training on the user's bladder. It can effectively perform urination tracking training on the user's bladder based on ultrasonic detection, can effectively improve the user's bladder control ability, and has a good effect on the user's bladder urination tracking training. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a module diagram of the bladder urination tracking training system based on ultrasonic detection of the present invention; Figure 2 It is a flowchart of the bladder urination tracking training system based on ultrasonic detection of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] In order to solve the problem that the existing technology cannot effectively perform urination tracking training on the user's bladder based on ultrasonic detection, cannot effectively improve the user's bladder control ability, and results in a poor effect on the user's bladder urination tracking training, please refer to Figure 1 - Figure 2 This embodiment provides the following technical solutions: A bladder urination tracking training system based on ultrasonic detection includes: an image acquisition module, an image processing module, a bladder evaluation module, and a urination training module.

[0019] Specifically, the image acquisition module acquires the real-time image of the user's bladder based on an ultrasound probe. The image processing module preprocesses, regionally segments, and extracts features from the acquired real-time image of the user's bladder to determine the user's bladder feature data. The bladder evaluation module constructs a bladder function evaluation model to analyze and perform biometric recognition on the user's bladder feature data to determine the user's bladder function evaluation result. The urination training module formulates a urination training plan for the user's bladder and conducts urination tracking training on the user's bladder. Therefore, through the interaction among the image acquisition module, the image processing module, the bladder evaluation module, and the urination training module, effective urination tracking training can be performed on the user's bladder based on ultrasonic detection, which can effectively improve the user's bladder control ability and achieve good results in the user's bladder urination tracking training.

[0020] Among them, the image acquisition module is used to acquire the real-time image of the user's bladder; In this embodiment, acquiring the real-time image of the user's bladder includes: Placing the ultrasound probe of a medical ultrasonic instrument on the lower abdomen of the user, and performing real-time detection and acquisition of the filling degree, shape, and structure of the user's bladder through the ultrasound probe to obtain the real-time image of the user's bladder based on ultrasonic detection.

[0021] Specifically, the filling degree, shape, and structure of the user's bladder are detected and acquired in real time through the ultrasound probe. It uses high-frequency sound waves to penetrate human tissues, and the reflected sound wave signals are processed by a computer to generate the real-time image of the user's bladder, thereby obtaining information such as the filling degree, shape, and structure of the bladder.

[0022] Among them, the image processing module is used to preprocess, regionally segment, and extract features from the acquired real-time image of the user's bladder to determine the user's bladder feature data; In this embodiment, preprocessing the acquired real-time image of the user's bladder includes: Denosing the real-time image of the user's bladder based on the Gaussian filtering algorithm, removing the speckle noise in the real-time image of the user's bladder, retaining the user's bladder boundary information in the real-time image of the user's bladder, and reducing the noise interference in the real-time image of the user's bladder; Enhancing the contrast of the real-time image of the user's bladder based on histogram equalization, highlighting the user's bladder boundary in the real-time image of the user's bladder, and making the user's bladder area in the real-time image of the user's bladder clear; Normalizing the real-time image of the user's bladder. Among them, performing gray-scale normalization on the real-time image of the user's bladder, scaling the pixel values in the real-time image of the user's bladder to a fixed range, and adjusting the size of the real-time image of the user's bladder to unify the resolution and eliminate the dimensional difference in the real-time image of the user's bladder.

[0023] In this embodiment, regionally segmenting and extracting features from the acquired real-time image of the user's bladder includes: Based on the threshold segmentation method, set a threshold according to the gray value of the user's real-time bladder image, and perform regional segmentation on the user's real-time bladder image according to the threshold to separate the bladder region; Extract features from the segmented bladder region, extract features useful for bladder urination tracking training from the segmented bladder region, and determine the user's bladder feature data, including bladder volume characteristics, bladder wall thickness, residual urine volume, bladder shape characteristics, bladder wall movement characteristics, bladder internal echo characteristics, and bladder pressure characteristics.

[0024] Specifically, the bladder volume characteristic is the volume change characteristic of the urine in the bladder, which reflects the filling degree of the bladder, is used to judge whether the bladder is full, guide urination training, monitor the change of bladder volume, and evaluate the training effect; the bladder wall thickness is the average thickness of the bladder wall, which reflects the health status of the bladder wall and is used to evaluate the elastic function of the bladder wall; the residual urine volume is the amount of urine remaining in the bladder after urination and is used to evaluate the urination efficiency and training effect; the bladder shape characteristic is the geometric shape characteristic of the bladder, such as symmetry, regularity, etc., and is used to evaluate the morphological change of the bladder; the bladder wall movement characteristic is the movement characteristic of the bladder wall during filling and urination, and is used to evaluate the contraction function and coordination of the bladder wall; the bladder internal echo characteristic is the echo characteristic of the urine in the bladder, which reflects the urine composition or abnormal structure and is used to assist in diagnosing bladder internal lesions; the bladder pressure characteristic is the change characteristic of the bladder internal pressure and is used to evaluate the contraction function and compliance of the bladder.

[0025] In this embodiment, extract features useful for bladder urination tracking training from the segmented bladder region, including: Extract texture and shape features from the segmented bladder region to obtain texture features and shape features; Determine key point information from the preset standard bladder feature information, and extract features from the bladder region based on the key point information to obtain key point features; Perform feature clustering on the texture features, shape features, and key point features respectively, and based on the feature clustering results, obtain target texture features, target shape features, and target key point features; Based on the bladder feature indicators, perform targeted feature fusion on the target texture features, target shape features, and target key point features to obtain the initial bladder features corresponding to each bladder feature indicator; Based on the standard key display feature information of each bladder feature indicator, verify the initial bladder features, and judge whether the clarity of the standard key display feature information in the initial bladder features is greater than the preset clarity, and whether the proportion of the standard key display feature information in the initial bladder features is greater than the preset proportion; If so, determine that the initial bladder features pass the verification, and use the initial bladder features as the user's bladder feature data; Otherwise, when the clarity of the standard key display feature information in the initial bladder feature is not greater than the preset clarity, feature enhancement processing is performed on the standard key display feature information in the initial bladder feature, and when the proportion of the standard key display feature information in the initial bladder feature is not greater than the preset proportion, other information in the initial bladder feature is eliminated to obtain the target bladder feature, and the target bladder feature is used as the user bladder feature data.

[0026] In this embodiment, the bladder characteristic indicators are bladder volume characteristics, bladder wall thickness, residual urine volume, bladder shape characteristics, bladder wall motion characteristics, intra-bladder echo characteristics and bladder pressure characteristics.

[0027] In this embodiment, the key point information determined in the preset standard bladder characteristic information is set based on historical experience.

[0028] In this embodiment, the purpose of performing feature clustering on the texture features, shape features and key point features is to extract similar features and improve the accuracy of image recognition.

[0029] In this embodiment, feature enhancement processing is performed on the standard key display feature information in the initial bladder feature in order to improve feature differentiation capability, and other information in the initial bladder feature is eliminated in order to avoid redundancy of feature information.

[0030] The beneficial effects of the above design scheme are: by extracting features of the bladder area from three aspects of texture, shape and key points, the comprehensiveness of the extracted features is ensured; by performing targeted feature fusion on the target texture features, target shape features and target key point features based on bladder feature indicators, the initial bladder features corresponding to each bladder feature indicator are obtained, the targetedness of the obtained initial bladder features to the feature indicators is ensured, and the initial bladder features are verified and optimized from two aspects of clarity and information ratio, the accuracy of the user bladder feature data finally obtained is ensured, and an accurate feature basis is provided for user bladder urination tracking training.

[0031] The bladder assessment module is used to construct a bladder function assessment model to analyze the user's bladder feature data and perform biometric identification to determine the user's bladder function assessment results; In this embodiment, a bladder function assessment model is constructed, including: According to the bladder urination tracking training requirements, bladder history data is collected and divided into a training set and a test set; Based on deep learning technology, a training set is used to train the deep learning model, so that the deep learning model can autonomously learn bladder function assessment behavior from the training set, and determine a bladder function assessment model based on deep learning; Based on the cross-validation method, a test set is used to perform performance testing on the deep learning-based bladder function evaluation model, and it is evaluated whether the deep learning-based bladder function evaluation model can achieve the expected effect of evaluating bladder function; According to the model test evaluation results, the parameters of the deep learning-based bladder function evaluation model are adjusted, and the deep learning-based bladder function evaluation model after parameter adjustment is continuously optimized until the deep learning-based bladder function evaluation model can achieve the expected effect of evaluating bladder function, and the best bladder function evaluation model is determined.

[0032] In this embodiment, the user's bladder characteristic data is analyzed and biometric identification is performed, including: Deploy the best bladder function evaluation model to the actual bladder function evaluation environment; Input the user's bladder characteristic data into the bladder function evaluation model, analyze and perform biometric identification on the user's bladder characteristic data according to the bladder function evaluation model, and evaluate the user's bladder function to determine the user's bladder function evaluation result, including the urine storage and urination functions of the user's bladder.

[0033] Among them, the urination training module is used to formulate a user's bladder urination training plan and perform urination tracking training on the user's bladder.

[0034] In this embodiment, performing urination tracking training on the user's bladder includes: Formulate a user's bladder urination training plan according to the user's bladder function evaluation result; Perform urination training on the user's bladder according to the formulated user's bladder urination training plan; Among them, the process of performing urination training on the user's bladder is monitored in real time, the user's feedback is recorded, when abnormal behavior is found, timely warning is given, and the effect of the user's bladder urination training is tracked in real time, the effect of the user's bladder urination training plan is evaluated, and the user's bladder urination training plan is adjusted according to the evaluation effect and the user's feedback to optimize the user's bladder urination training plan.

[0035] In this embodiment, performing urination training on the user's bladder according to the formulated user's bladder urination training plan includes: According to the user's bladder function evaluation result, formulate a scientific drinking plan for the user, including: the total daily water intake, the number of drinking times, the drinking time, and the single water intake, to avoid overfilling the bladder due to drinking a large amount of water in a short period of time; Guide the user to perform contraction and relaxation exercises of the pelvic floor muscles, each time lasting for 3 - 5 seconds, perform 3 - 5 groups per day, and each group has 10 - 15 times, and monitor the contraction of the user's pelvic floor muscles in real time, and adjust the pelvic floor muscle exercise according to the monitoring situation in a timely manner; Guide the user to urinate according to a fixed schedule, gradually extend the urination interval, and stimulate the urination reflex by the sound of running water and applying hot compress to the lower abdomen, so as to enable the user to establish the awareness of autonomous urination.

[0036] In this embodiment, according to the model test evaluation results, the parameters of the bladder function evaluation model based on deep learning are adjusted, including: Combined with the model test evaluation results, intelligently generate a grid including multiple parameter combinations of learning rate, number of layers, and number of neurons, and obtain the evaluation index under each model parameter combination based on grid search; Based on the evaluation index, calculate the comprehensive performance evaluation value of each model parameter combination; Among them, represents the comprehensive performance evaluation value of the current model parameter combination, represents the accuracy of the model under the current model parameter combination, represents the accuracy threshold, represents the precision of the model under the current model parameter combination, represents the precision threshold, represents the recall rate of the model under the current model parameter combination, represents the recall rate threshold, represents the accuracy weight, represents the precision weight, represents the recall rate weight; Based on the comprehensive performance evaluation value, combined with the model complexity under the model parameter combination, calculate the weighted weight of each model parameter combination; Among them, represents the weighted weight of the current model parameter combination, represents the natural constant, with a value of 2.72, represents the model complexity under the current model parameter combination; Select the target model parameter combination with a weighted weight greater than the preset weight, and perform weighted averaging on the prediction results under the target model parameter combination based on the weighted weight to obtain the final model prediction result.

[0037] In this embodiment, the evaluation indexes include accuracy, precision, and recall rate. The thresholds of accuracy, precision, and recall rate are flexibly set according to the actual situation, and the weights of accuracy, precision, and recall rate are set according to the requirements of the model.

[0038] In this embodiment, the model complexity is related to the number of layers, the number of nodes, etc.

[0039] In this embodiment, the higher the comprehensive performance evaluation value of the model parameter combination, the higher the corresponding weighted weight of the combination.

[0040] In this embodiment, the higher the model complexity, the lower the corresponding weighted weight of the combination, aiming to ensure the efficiency and accuracy during model fusion.

[0041] The beneficial effects of the above design scheme are as follows: By means of grid search, multiple model parameter combinations are determined, and the comprehensive performance evaluation value of each model parameter combination is analyzed. Based on the comprehensive performance evaluation value and combined with the model complexity under the model parameter combination, the weighted weight of each model parameter combination is calculated. The larger the weighted weight, the larger the proportion during fusion. By training and fusing multiple models, the prediction results of each model are weighted and averaged, leveraging the advantages under different model parameter combinations to improve the stability and generalization ability of the model, providing a high-quality model basis for further bladder detection based on the model.

[0042] In summary, by using an ultrasonic probe to perform real-time detection and acquisition of the filling degree, shape, and structure of the user's bladder, obtaining real-time images of the user's bladder based on ultrasonic detection, preprocessing, region segmentation, and feature extraction of the acquired real-time images of the user's bladder to determine the user's bladder feature data, constructing a bladder function evaluation model according to the bladder urination tracking training requirements, analyzing and biometrically identifying the user's bladder feature data based on the bladder function evaluation model, and evaluating the user's bladder function to determine the user's bladder function evaluation result, formulating a user's bladder urination training plan according to the user's bladder function evaluation result and performing urination tracking training on the user's bladder, effective urination tracking training of the user's bladder can be carried out based on ultrasonic detection, effectively improving the user's bladder control ability and achieving good results in the user's bladder urination tracking training.

[0043] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0044] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A bladder urination tracking and training system based on ultrasonic detection, characterized in that, Including: An image acquisition module configured to acquire real-time images of a user's bladder based on an ultrasound probe; An image processing module configured to preprocess, regionally segment, and extract features from the acquired real-time images of the user's bladder to determine user bladder feature data; Based on the standard key display feature information of each bladder feature index, verify the initial bladder features, determine whether the clarity of the standard key display feature information in the initial bladder features is greater than a preset clarity, and whether the proportion of the standard key display feature information in the initial bladder features is greater than a preset proportion, so as to determine the user bladder feature data; A bladder evaluation module configured to construct a bladder function evaluation model to analyze and biometrically identify the user bladder feature data to determine the user bladder function evaluation result; A urination training module configured to formulate a user bladder urination training plan according to the user bladder function evaluation result and perform urination tracking training on the user's bladder.

2. The bladder urination tracking and training system based on ultrasonic detection according to claim 1, wherein Acquiring real-time images of the user's bladder, including: Placing the ultrasound probe of a medical ultrasound instrument on the lower abdomen of the user, and performing real-time detection and acquisition of the fullness, shape, and structure of the user's bladder through the ultrasound probe to obtain real-time images of the user's bladder based on ultrasound detection.

3. The bladder urination tracking and training system based on ultrasonic detection according to claim 1, characterized in that Preprocessing the acquired real-time images of the user's bladder, including: Denosing the real-time images of the user's bladder based on the Gaussian filtering algorithm, removing the speckle noise in the real-time images of the user's bladder, retaining the user bladder boundary information in the real-time images of the user's bladder, and reducing the noise interference in the real-time images of the user's bladder; Enhancing the contrast of the real-time images of the user's bladder based on histogram equalization, highlighting the user bladder boundary in the real-time images of the user's bladder, and making the user bladder area in the real-time images of the user's bladder clear; Normalizing the real-time images of the user's bladder, wherein, performing gray normalization on the real-time images of the user's bladder, scaling the pixel values in the real-time images of the user's bladder to a fixed range, and adjusting the size of the real-time images of the user's bladder to unify the resolution and eliminate the dimension difference in the real-time images of the user's bladder.

4. The bladder urination tracking and training system based on ultrasonic detection according to claim 1, wherein Performing regional segmentation and feature extraction on the acquired real-time images of the user's bladder, including: Based on the threshold segmentation method, setting a threshold according to the gray value of the real-time images of the user's bladder, and performing regional segmentation on the real-time images of the user's bladder according to the threshold to separate the bladder region; Performing feature extraction on the segmented bladder region, extracting features useful for bladder urination tracking training from the segmented bladder region, and determining user bladder feature data, including bladder volume features, bladder wall thickness, residual urine volume, bladder shape features, bladder wall movement features, bladder internal echo features, and bladder pressure features.

5. The bladder urination tracking and training system based on ultrasonic detection according to claim 4, characterized in that, Extracting features useful for bladder urination tracking training from the segmented bladder region, including: Performing feature extraction on the segmented bladder region from texture and shape to obtain texture features and shape features; Determining key point information from the preset standard bladder feature information, and performing feature extraction on the bladder region based on the key point information to obtain key point features; Performing feature clustering on the texture features, shape features, and key point features respectively, and based on the feature clustering results, obtaining target texture features, target shape features, and target key point features; Based on the bladder feature index, the target texture feature, the target shape feature and the target key point feature are subjected to targeted feature fusion to obtain an initial bladder feature corresponding to each bladder feature index; Based on the standard key display feature information of each bladder feature indicator, the initial bladder feature is verified to determine whether the clarity of the standard key display feature information in the initial bladder feature is greater than the preset clarity, and whether the proportion of the standard key display feature information in the initial bladder feature is greater than the preset proportion; If yes, determining that the initial bladder feature verification is passed, and using the initial bladder feature as user bladder feature data; Otherwise, when the clarity of the standard key display feature information in the initial bladder feature is not greater than the preset clarity, feature enhancement processing is performed on the standard key display feature information in the initial bladder feature; when the proportion of the standard key display feature information in the initial bladder feature is not greater than the preset proportion, other information in the initial bladder feature is eliminated to obtain the target bladder feature, and the target bladder feature is used as the user bladder feature data.

6. The bladder urination tracking and training system based on ultrasonic detection according to claim 1, wherein Construct a bladder function assessment model, including: According to the bladder urination tracking training requirements, bladder history data is collected and divided into a training set and a test set; Based on deep learning technology, a training set is used to train the deep learning model, so that the deep learning model can autonomously learn bladder function assessment behavior from the training set, and determine a bladder function assessment model based on deep learning; Based on the cross-validation method, the test set was used to test the performance of the deep learning-based bladder function assessment model, and to evaluate whether the deep learning-based bladder function assessment model can achieve the expected effect of assessing bladder function; According to the model test evaluation results, the parameters of the deep learning-based bladder function evaluation model are adjusted, and the deep learning-based bladder function evaluation model after parameter adjustment is continuously optimized until the deep learning-based bladder function evaluation model can achieve the expected effect of evaluating bladder function and determine the optimal bladder function evaluation model.

7. The bladder urination tracking and training system based on ultrasonic detection according to claim 6, characterized in that Analyze the user's bladder characteristic data and identify biometric characteristics, including: Deploy the best bladder function assessment model in the actual bladder function assessment environment; The user's bladder characteristic data is input into a bladder function evaluation model, and the user's bladder characteristic data is analyzed and biometrically identified according to the bladder function evaluation model, and the user's bladder function is evaluated to determine the user's bladder function evaluation results, including the urine storage and urination functions of the user's bladder.

8. The bladder urination tracking and training system based on ultrasonic detection according to claim 1, characterized in that The user's bladder is trained to track urination, including: Develop a bladder urination training program for users based on the user's bladder function assessment results; Performing urination training on the user's bladder according to a developed bladder urination training program for the user; Among them, the process of user bladder urination training is monitored in real time, user feedback is recorded, and when abnormal behavior is found, timely warning is given. The effect of the user's bladder urination training is tracked in real time, and the effect of the user's bladder urination training plan is evaluated. The user's bladder urination training plan is adjusted according to the evaluation effect and user feedback to optimize the user's bladder urination training plan.

9. The bladder urination tracking and training system based on ultrasonic detection according to claim 8, wherein Perform bladder urination training on the user's bladder according to the formulated user bladder urination training plan, including: Based on the user's bladder function assessment results, formulate a scientific drinking plan for the user, including: total daily water intake, drinking frequency, drinking time, and single water intake; Guide the user to perform contraction and relaxation exercises of the pelvic floor muscles, each time lasting 3-5 seconds, 3-5 groups per day, 10-15 times per group, and monitor the contraction of the user's pelvic floor muscles in real time, and adjust the pelvic floor muscle exercise according to the monitoring situation in a timely manner; Guide the user to urinate according to a fixed schedule, gradually extend the urination interval, and stimulate the urination reflex by the sound of running water and hot compress on the lower abdomen to enable the user to establish the awareness of autonomous urination.

10. The bladder urination tracking and training system based on ultrasonic detection according to claim 6, characterized in that, According to the model test evaluation results, adjust the parameters of the bladder function evaluation model based on deep learning, including: Combined with the model test evaluation results, intelligently generate a grid including multiple parameter combinations of learning rate, number of layers, and number of neurons, and obtain the evaluation index under each model parameter combination based on grid search; Based on the evaluation index, calculate the comprehensive performance evaluation value of each model parameter combination; Based on the comprehensive performance evaluation value, combined with the model complexity under the model parameter combination, calculate the weighted weight of each model parameter combination; Select the target model parameter combination with a weighted weight greater than the preset weight, and perform weighted averaging on the prediction results under the target model parameter combination based on the weighted weight to obtain the final model prediction result.

Images