Transesophageal visual intelligent ultrasonic probe

Abstract

The invention discloses a transesophageal visual intelligent ultrasonic probe, and belongs to the technical field of medical instrument image processing. The combination of the sliding rail, the sliding block and the locking switch is adopted, the supporting plate can freely slide and be stably locked, different operation requirements are met, it is guaranteed that the probe is stable, rapid fixing can be achieved when the probe is not used, the burden of medical staff is relieved, the ultrasonic image and the physiological parameters of a patient are combined, and the ultrasonic image and the physiological parameters of the patient are combined through feature matching, data association and fusion calculation. A cross-modal intelligent analysis system is established, the limitation of single ultrasonic image diagnosis is broken through, more accurate lesion recognition and diagnosis support is provided, accurate segmentation and feature extraction are performed on a lesion area through deep learning, lesion details are enhanced through histogram equalization, meanwhile, a lesion contour is automatically drawn, a disease type and related information are marked, and the accuracy of lesion recognition and diagnosis is improved. The ultrasonic image is more visual, and the diagnosis efficiency and accuracy of doctors are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical device image processing, and in particular to a transesophageal visualization intelligent ultrasound probe. Background Art

[0002] Ultrasound examination, as a non-invasive, real-time, dynamic imaging medical imaging technology, is widely used in the fields of cardiovascular, digestive system and intensive care. Among them, transesophageal ultrasound (TEE) has become an important tool in cardiology, anesthesiology and perioperative monitoring because it can be close to the heart and large blood vessels and provide high-resolution anatomical and blood flow information. However, existing transesophageal ultrasound probes and imaging systems still have the following problems: Traditional TEE probes are mainly operated by medical staff, which can easily lead to hand fatigue after long-term use, affecting the stability and accuracy of ultrasound examinations. At the same time, some existing bracket structures have a single fixing method, which is difficult to flexibly adjust according to the body shape or examination needs of different patients, which is not conducive to optimizing the operating experience.

[0003] Existing TEE imaging systems mainly rely on ultrasound images to judge lesions, and lack comprehensive analysis of patients' physiological parameters such as heart rate, blood pressure, and blood oxygen. As a result, diagnosis is based only on imaging features, which may miss potential pathological information and affect the accurate identification and early intervention of the disease.

[0004] Traditional ultrasound images rely on doctors’ subjective experience for interpretation and lack intelligent analysis and automatic annotation functions. Especially in low-contrast or complex lesions, the lesion area is difficult to identify, which may lead to misjudgment. At the same time, doctors need to manually measure the size of the lesion and analyze the echo characteristics, which increases the diagnosis time and burden. Summary of the invention

[0005] The purpose of the present invention is to provide a transesophageal visualization intelligent ultrasound probe to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a transesophageal visualized intelligent ultrasound probe, comprising an ultrasound probe and a main body bracket, characterized in that: a slide rail is provided on the main body bracket, a support plate is slidably provided on the slide rail, probe handle mounting sleeves are provided on both sides of the support plate, and the ultrasound probe can be clamped in the probe handle mounting sleeve; The ultrasound probe is equipped with an imaging system for manipulation, and the imaging system is configured to reconstruct ultrasound images and generate comprehensive analysis results and diagnostic conclusions. When generating the diagnostic conclusion, the disease matching degree is obtained by combining the influencing factors of the disease characteristics with the weights of the pathological characteristics and the feature similarity to determine the disease type of the lesion area.

[0007] Furthermore, side pull rods are connected to both sides of the probe handle mounting sleeve, a pulling shaft is vertically arranged on the top of the side pull rod, and the top axle pin of the pulling shaft is connected to the linkage shaft. Two clamping plates are arranged above the probe handle mounting sleeve, and the middle part of the outer side of the clamping plate is connected to the linkage shaft through the axle pin, and a guide strip is arranged on the outer side of the clamping plate. Guide plates are arranged on both sides of the probe handle mounting sleeve, and the guide strip slides through the guide plate. A reset spring is arranged at the bottom of the side pull rod.

[0008] Furthermore, a slide groove is provided on the slide rail, a sliding block is slidably arranged in the slide groove, a locking switch is provided on the sliding block, the bottom surface of the support plate is connected to the top of the sliding block through an axle pin, a support foot is provided at the bottom of the sliding block, and a groove matching the support foot is provided on the bottom surface of the support plate. A folding frame is provided on one side of the main bracket, and a connecting plate is provided on the side of the folding frame away from the main bracket, a first chuck and a second chuck are provided on the connecting plate, and the ultrasonic probe can be clamped in the probe handle mounting sleeve, the first chuck or the second chuck.

[0009] Furthermore, a transesophageal visualization intelligent imaging system is applied to the above-mentioned transesophageal visualization intelligent ultrasound probe, including: an ultrasound signal acquisition module, an intelligent image reconstruction module, a multimodal fusion analysis module, a disease intelligent diagnosis module and an image optimization and annotation module; The ultrasonic signal acquisition module is configured to acquire ultrasonic echo signals through an ultrasonic probe passing through the esophagus and convert them into digital signals; The intelligent image reconstruction module is configured to extract features of the ultrasonic echo signal after obtaining the ultrasonic echo signal, and reconstruct a high-resolution ultrasonic image; The multimodal fusion analysis module is configured to perform fusion analysis on the reconstructed ultrasound image and the patient's physiological parameter data to generate a comprehensive analysis result; The disease intelligent diagnosis module is configured to identify and judge the lesion characteristics in the ultrasound image based on the comprehensive analysis results and in combination with the disease characteristic knowledge base, and output a diagnosis conclusion; The image optimization and annotation module is configured to perform targeted optimization processing on the ultrasound image according to the diagnosis conclusion, enhance the display effect of the lesion area, add annotation information related to the diagnosis result, and generate an annotated image.

[0010] Furthermore, the multimodal fusion analysis module includes: A feature matching unit is configured to extract image features layer by layer from the reconstructed ultrasound image, wherein the image features include edge contours, grayscale distribution, and texture features, model the patient's physiological parameter data using a time series and extract physiological parameter time series features, wherein the physiological parameter time series features include trend changes, extreme points, and periodicity, match the image features with the physiological parameter time series features, calculate a correlation score, and screen a feature combination; A data association unit, configured to convert image features and physiological parameter time series features into high-dimensional feature vectors, construct a graph structure between lesion features, image features and physiological parameter time series features, and define association weights of nodes and edges of the graph structure; The fusion calculation unit is configured to calculate a final fusion feature vector based on correlation scores of different modal features, mark abnormal data based on the fusion feature vector, and generate a comprehensive analysis result.

[0011] Furthermore, the disease intelligent diagnosis module includes: A lesion recognition unit is configured to perform pixel-level segmentation on the ultrasound image, extract possible lesion areas, perform secondary feature extraction on the lesion areas, obtain lesion features, wherein the lesion features include boundary morphology, signal intensity, and internal texture, and use a deep neural network to classify the lesion features into lesion types; The diagnostic reasoning unit is configured to match the lesion features with the disease feature knowledge base and calculate the matching degree, determine the disease type of the lesion area, perform comprehensive reasoning based on rule reasoning combined with multiple lesion features, and generate a diagnostic conclusion.

[0012] Furthermore, the image optimization and annotation module includes: an adaptive enhancement unit configured to use histogram equalization to improve the visibility of image details in low-contrast lesion areas, enhance texture information in the lesion areas, and perform edge enhancement on the lesion areas of the ultrasound image; The intelligent labeling unit is configured to draw the lesion outline on the ultrasound image and label the corresponding disease type according to the segmentation result of the lesion area, and superimpose text information related to the disease type on the ultrasound image, wherein the text information includes the lesion size, type and severity, and generate a labeled image.

[0013] Furthermore, the patient's physiological parameter data includes heart rate, blood pressure, blood oxygen saturation and respiratory rate, and the patient's physiological parameter data is pre-collected before the ultrasound image is collected.

[0014] Furthermore, the disease feature knowledge base includes morphological features, echo characteristics, hemodynamic parameters of different diseases on ultrasound images, and corresponding disease diagnosis basis, differential diagnosis information and treatment recommendations.

[0015] Furthermore, when the intelligent image reconstruction module extracts features from the ultrasonic echo signal, it includes: Performing signal preprocessing on the ultrasonic echo signal to obtain a first processed ultrasonic echo signal; Verify the first processed ultrasonic echo signal according to the time point, analyze whether the signals at the same time point are aligned, and obtain the verification analysis result; Performing alignment adjustment on the first processed ultrasonic echo signal according to the verification analysis result to obtain a second processed ultrasonic echo signal; Performing time domain feature analysis based on the second processed ultrasonic echo signal to obtain the time domain feature of the second processed ultrasonic echo signal and obtain a first feature of the ultrasonic echo signal; Converting the second processed ultrasonic echo signal from the time domain to the frequency domain, and performing frequency domain feature analysis to obtain a second feature of the ultrasonic echo signal; Decomposing the second processed ultrasonic echo signal, decomposing the second processed ultrasonic echo signal into a plurality of wavelet signals of different scales, and performing feature analysis on the wavelet signal to obtain a third feature of the ultrasonic echo signal; parsing the second processed ultrasonic echo signal, and analyzing the intensity change information of the second processed ultrasonic echo signal to obtain a fourth feature of the ultrasonic echo signal; Normalization analysis is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal, and normalization adjustment is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal to obtain the ultrasonic echo signal feature extraction result.

[0016] Furthermore, when the diagnostic reasoning unit matches the lesion feature with the disease feature knowledge base and calculates the matching degree, it includes: According to the pathological features, semantic similarity is calculated in the disease feature knowledge base according to the disease features to obtain feature similarity; Screening is performed based on the feature similarity combined with the first threshold, and disease features corresponding to feature similarities greater than the first threshold are screened out to obtain a first screening result; In the first screening results, the target disease is determined based on the disease characteristics; Determine the number of pathological features; When the number of pathological features is one, the disease features of the target disease are acquired to obtain the disease features of the target disease, and the disease influence analysis is performed on the disease features of the target disease to determine the influencing factors of the disease features, and the influencing factors of the disease features are combined with the feature similarity to obtain the disease matching degree; When the number of pathological features is two or more, the disease matching degree is calculated for the pathological features in combination with the target disease, and at the same time, the pathological features are analyzed to see whether there is any association, and the association analysis results are obtained; Combining the pathological features according to the association analysis results to determine the pathological feature combination results; The disease characteristics of the target disease are acquired to obtain the disease characteristics of the target disease, and the disease influence analysis is performed on the disease characteristics of the target disease to determine the influencing factors of the disease characteristics. At the same time, the pathological feature weight analysis is performed on the pathological feature combination results to determine the weight of the pathological feature. Then, the influencing factors of the disease characteristics are combined with the weight of the pathological characteristics and the feature similarity to obtain the disease matching degree.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts the design of slide rails, sliding blocks and locking switches, so that the support plate can be flexibly adjusted and firmly locked, providing a more convenient operation experience for medical staff. At the same time, the combination of the folding frame, the probe handle mounting sleeve and the chuck not only adapts to the physiological characteristics and examination needs of different patients, but also can achieve rapid fixation when the probe handle is not in use, reducing the burden on medical staff, improving the applicability, operation convenience and stability of the equipment, and meeting the different needs of routine examinations and special examinations.

[0018] 2. The present invention combines ultrasound images with patient physiological parameter data, establishes a cross-modal data analysis model through feature matching, data association and fusion calculation, extracts ultrasound image features and physiological parameter time series features, calculates correlation scores, screens the most relevant feature combinations, and uses graph structure modeling to clarify the relationship between lesion characteristics, ultrasound images and physiological parameters, improves the depth of data analysis, calculates the final fusion feature vector, accurately marks abnormal data, and generates comprehensive analysis results, breaking through the limitations of single ultrasound image diagnosis. Through cross-modal information fusion, it provides more comprehensive and accurate support for disease diagnosis and improves the clinical decision-making value of ultrasound examinations.

[0019] 3. The present invention integrates deep neural network lesion recognition and intelligent labeling functions to optimize the readability of ultrasound images. Through pixel-level segmentation, secondary feature extraction and deep learning, it can achieve high-precision lesion area recognition, help doctors discover early tiny lesions, combine with the disease feature knowledge base, calculate the lesion matching degree, and make a comprehensive judgment based on rule reasoning to improve the scientificity and reliability of diagnosis. The image optimization and labeling module improves the visibility of details in low-contrast areas, and automatically draws lesion contours, labels disease types, and superimposes diagnostic information, making the diagnostic results more intuitive and facilitating doctors to make quick decisions. It makes the visualization of ultrasound images more intuitive, the diagnosis more accurate, and the doctor's operation more efficient, greatly improving the clinical value and applicability of ultrasound examinations. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the placement structure of the ultrasonic probe of the present invention; Figure 2 It is a schematic diagram of the bottom structure of the main support of the present invention; Figure 3It is a schematic diagram of the bottom structure of the support plate of the present invention; Figure 4 This is a schematic diagram of the probe handle mounting sleeve connection structure of the present invention; Figure 5 It is a schematic diagram of the imaging system module of the present invention.

[0021] In the figure: 1. ultrasonic probe; 2. main bracket; 3. slide rail; 4. support plate; 5. folding frame; 6. connecting plate; 7. first chuck; 8. second chuck; 9. slide groove; 10. sliding block; 11. groove; 12. probe handle mounting sleeve; 13. locking switch; 14. support foot; 15. side opening; 16. side pull rod; 17. pulling shaft; 18. reset spring; 19. linkage shaft; 20. clamping plate; 21. inner cushion layer; 22. guide strip; 23. guide plate. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1-4 , the present invention provides the following technical solutions: The transesophageal visualization intelligent ultrasound probe comprises an ultrasound probe 1 and a main body support 2, wherein the main body support 2 is provided with a slide rail 3, a support plate 4 is slidably provided on the slide rail 3, and probe handle mounting sleeves 12 are provided on both sides of the support plate 4, and the ultrasound probe 1 can be clamped in the probe handle mounting sleeve 12; The probe handle mounting sleeve 12 is connected to side pull rods 16 on both sides, a pulling shaft 17 is vertically arranged on the top of the side pull rod 16, and the top axle pin of the pulling shaft 17 is connected to a linkage shaft 19, and two clamping plates 20 are arranged above the probe handle mounting sleeve 12, and the middle part of the outer side of the clamping plate 20 is connected to the linkage shaft 19 through an axle pin, and a guide bar 22 is arranged on the outer side of the clamping plate 20, and guide plates 23 are arranged on both sides of the probe handle mounting sleeve 12, and the guide bar 22 slides through the guide plate 23, and a reset spring 18 is arranged at the bottom of the side pull rod 16.

[0024] A chute 9 is provided on the sliding rail 3. A sliding block 10 is slidably arranged in the chute 9. A locking switch 13 is arranged on the sliding block 10. The bottom surface of the support plate 4 is connected to the top of the sliding block 10 through a pin. A support leg 14 is arranged at the bottom of the sliding block 10. A groove 11 matching the support leg 14 is provided on the bottom surface of the support plate 4. A folding frame 5 is arranged on one side of the main body bracket 2. A connecting plate 6 is arranged on the side of the folding frame 5 away from the main body bracket 2. A first chuck 7 and a second chuck 8 are arranged on the connecting plate 6. The ultrasonic probe 1 can be clamped in the probe handle mounting sleeve 12, the first chuck 7 or the second chuck 8.

[0025] In the above embodiment, when the handle of the ultrasonic probe 1 is inserted into the interior of the probe handle mounting sleeve 12 from top to bottom, the probe handle mounting sleeve 12 moves downward under the action of gravity, driving the side pull rod 16 to move downward horizontally, then pulling the pull shaft 17, and subsequently driving the linkage shaft 19 to move. At this time, the linkage shaft 19 horizontally pushes out, pushing the two clamping plates 20 towards each other to clamp the handle of the ultrasonic probe 1. When taking out the handle of the ultrasonic probe 1, only need to pull it upward, and the return spring 18 drives the probe handle mounting sleeve 12 to reset upward. At this time, the linkage shaft 19 pulls the pull shaft 17 outward under the action of the reset of the probe handle mounting sleeve 12, thereby driving the two clamping plates 20 to move in the opposite direction, completing the automatic locking and release of the handle of the ultrasonic probe 1.

[0026] In the above embodiment, the design of the sliding rail 3 and the sliding block 10, in cooperation with the locking switch 13, enables the support plate 4 to flexibly adjust its position and be firmly locked, facilitating the operation of medical staff and placing items. The setting of the folding frame 5 and the connecting plate 6, as well as multiple components for fixing the ultrasonic probe 1, such as the probe handle mounting sleeve 12, the first chuck 7 and the second chuck 8, provide multiple installation and fixing methods for the handle of the ultrasonic probe 1. This adjustability not only improves the convenience of operation but also can adapt to the physiological characteristics and examination requirements of different patients, enhancing the applicability of the system. Different installation methods can meet the usage requirements in different scenarios. For example, during routine examinations and special examinations, medical staff can select the most suitable fixing method according to the actual situation to ensure the stable operation of the ultrasonic probe 1, thereby improving the accuracy and reliability of ultrasonic examinations. At the same time, it is convenient to fix the handle of the ultrasonic probe 1 when it is not needed during the operation, liberating the medical staff.

[0027] Please refer to Figure 5 , an imaging system for manipulation is configured in the ultrasonic probe 1. The imaging system is configured to reconstruct ultrasonic images and generate comprehensive analysis results and diagnostic conclusions. When generating a diagnostic conclusion, the disease matching degree is obtained by combining the influencing factors of disease characteristics with the weight values of pathological characteristics and the feature similarity, and the disease type of the lesion area is determined; The imaging system includes: an ultrasound signal acquisition module, an intelligent image reconstruction module, a multimodal fusion analysis module, a disease intelligent diagnosis module and an image optimization and annotation module; The ultrasonic signal acquisition module is configured to acquire ultrasonic echo signals through the transesophageal ultrasonic probe 1 and convert them into digital signals; The intelligent image reconstruction module is configured to extract features of the ultrasonic echo signal after obtaining the ultrasonic echo signal, and reconstruct a high-resolution ultrasonic image; The multimodal fusion analysis module is configured to perform fusion analysis on the reconstructed ultrasound image and the patient's physiological parameter data to generate a comprehensive analysis result; The disease intelligent diagnosis module is configured to identify and judge the lesion characteristics in the ultrasound image based on the comprehensive analysis results and in combination with the disease characteristic knowledge base, and output a diagnosis conclusion; The image optimization and annotation module is configured to perform targeted optimization processing on the ultrasound image according to the diagnosis conclusion, enhance the display effect of the lesion area, add annotation information related to the diagnosis result, and generate an annotated image.

[0028] In the above embodiment, the modules of the imaging system work closely together. The ultrasonic signal acquisition module can efficiently acquire ultrasonic echo signals and accurately convert them into digital signals. The intelligent image reconstruction module reconstructs high-resolution ultrasonic images by extracting features from ultrasonic echo signals. The multimodal fusion analysis module fuses the ultrasonic image with the patient's physiological parameter data, comprehensively analyzes the condition from multiple dimensions, and mines the potential connection between image features and physiological parameter time series features through feature matching, data association and fusion calculation, providing a more comprehensive and accurate basis for disease diagnosis. This multimodal fusion method breaks the limitations of single data, can reveal the nature of the disease more deeply, assist doctors in making more scientific diagnostic decisions, and improve the accuracy and reliability of diagnosis.

[0029] Multimodal fusion analysis module, including: A feature matching unit is configured to extract image features layer by layer from the reconstructed ultrasound image, wherein the image features include edge contours, grayscale distribution, and texture features, model the patient's physiological parameter data using a time series and extract physiological parameter time series features, wherein the physiological parameter time series features include trend changes, extreme points, and periodicity, match the image features with the physiological parameter time series features, calculate a correlation score, and screen a feature combination; A data association unit, configured to convert image features and physiological parameter time series features into high-dimensional feature vectors, construct a graph structure between lesion features, image features and physiological parameter time series features, and define association weights of nodes and edges of the graph structure; The fusion calculation unit is configured to calculate a final fusion feature vector based on correlation scores of different modal features, mark abnormal data based on the fusion feature vector, and generate a comprehensive analysis result.

[0030] In the above embodiment, the feature matching unit performs fine feature extraction on the ultrasound image and the patient's physiological parameter data, calculates the correlation score between them, and screens out valuable feature combinations, which can discover the potential correlation between different modal data, help extract key information from complex data, and provide strong support for subsequent diagnosis. The data association unit converts the features of different modalities into high-dimensional feature vectors, constructs a graph structure and defines association weights, so that the relationship between the data is clearer and quantifiable, which is convenient for computer analysis and processing, and improves the efficiency and accuracy of data processing. The fusion calculation unit calculates the fusion feature vector based on the correlation score, marks the abnormal data, and generates a comprehensive analysis result, which can comprehensively and accurately reflect the patient's condition, provide doctors with more valuable diagnostic information for reference, and help doctors judge the condition more accurately.

[0031] Disease intelligent diagnosis module, including: A lesion recognition unit is configured to perform pixel-level segmentation on the ultrasound image, extract possible lesion areas, perform secondary feature extraction on the lesion areas, obtain lesion features, wherein the lesion features include boundary morphology, signal intensity, and internal texture, and use a deep neural network to classify the lesion features into lesion types; The diagnostic reasoning unit is configured to match the lesion features with the disease feature knowledge base and calculate the matching degree, determine the disease type of the lesion area, perform comprehensive reasoning based on rule reasoning combined with multiple lesion features, and generate a diagnostic conclusion.

[0032] In the above embodiment, the lesion recognition unit can accurately locate the lesion area and extract key lesion features through pixel-level segmentation and secondary feature extraction, and adopts deep neural network to classify the lesion type, making full use of the powerful pattern recognition ability of deep learning, improving the accuracy of lesion type judgment, and helping doctors to detect tiny lesions early, providing the possibility for early treatment of the disease. The diagnostic reasoning unit matches and calculates the lesion features with the disease feature knowledge base, and combines rule reasoning for comprehensive reasoning, which can make full use of existing medical knowledge and clinical experience to draw accurate diagnostic conclusions, avoid the limitations of a single diagnostic method, improve the reliability and scientificity of the diagnosis, and provide a strong guarantee for the precise treatment of patients.

[0033] Image optimization and annotation module, including: an adaptive enhancement unit configured to use histogram equalization to improve the visibility of image details in low-contrast lesion areas, enhance texture information in the lesion areas, and perform edge enhancement on the lesion areas of the ultrasound image; The intelligent labeling unit is configured to draw the lesion outline on the ultrasound image and label the corresponding disease type according to the segmentation result of the lesion area, and superimpose text information related to the disease type on the ultrasound image, wherein the text information includes the lesion size, type and severity, and generate a labeled image.

[0034] In the above embodiment, the adaptive enhancement unit adopts histogram equalization and other technologies to improve the visibility of image details in low-contrast lesion areas, enhance texture information and edge enhancement effects, so that doctors can more clearly observe the morphology, boundaries and internal structure of the lesions, which helps to accurately judge the nature and severity of the lesions. The intelligent labeling unit draws the lesion outline, labels the disease type and superimposes relevant text information on the ultrasound image based on the segmentation results of the lesion area, providing doctors with concise and clear diagnostic prompts. These labeled information can help doctors quickly understand the key information of the lesions, reduce the information search and analysis time in the diagnosis process, improve the diagnosis efficiency, and also facilitate communication and exchanges between different doctors, which helps to improve the overall medical diagnosis level.

[0035] When the intelligent image reconstruction module extracts features from the ultrasonic echo signal, it includes: Performing signal preprocessing on the ultrasonic echo signal to obtain a first processed ultrasonic echo signal; Verify the first processed ultrasonic echo signal according to the time point, analyze whether the signals at the same time point are aligned, and obtain the verification analysis result; Performing alignment adjustment on the first processed ultrasonic echo signal according to the verification analysis result to obtain a second processed ultrasonic echo signal; Performing time domain feature analysis based on the second processed ultrasonic echo signal to obtain the time domain feature of the second processed ultrasonic echo signal and obtain a first feature of the ultrasonic echo signal; Converting the second processed ultrasonic echo signal from the time domain to the frequency domain, and performing frequency domain feature analysis to obtain a second feature of the ultrasonic echo signal; Decomposing the second processed ultrasonic echo signal, decomposing the second processed ultrasonic echo signal into a plurality of wavelet signals of different scales, and performing feature analysis on the wavelet signal to obtain a third feature of the ultrasonic echo signal; parsing the second processed ultrasonic echo signal, and analyzing the intensity change information of the second processed ultrasonic echo signal to obtain a fourth feature of the ultrasonic echo signal; Normalization analysis is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal, and normalization adjustment is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal to obtain the ultrasonic echo signal feature extraction result.

[0036] Among them, signal preprocessing includes: denoising processing, enhancement processing, etc.

[0037] If the verification and analysis result shows that the signals at the same time point are aligned, an alignment adjustment is performed on the first processed ultrasonic echo signal to obtain a second processed ultrasonic echo signal. If the verification and analysis result shows that the signals at the same time point are not aligned, the first processed ultrasonic echo signal is the second processed ultrasonic echo signal.

[0038] In the above embodiment, when the intelligent image reconstruction module extracts features from the ultrasonic echo signal, the ultrasonic echo signal is optimized by signal preprocessing, the signal-to-noise ratio of the ultrasonic echo signal is improved, the signal features are enhanced, and convenience is provided for feature extraction, so that feature extraction can be performed better. Moreover, by verifying the first processed ultrasonic echo signal according to the time point, the signal time alignment is ensured, convenience is provided for subsequent feature analysis, the confusion and error probability of feature analysis are reduced, and the accuracy of signal features is guaranteed. Moreover, by performing normalization analysis on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal, and the fourth feature of the ultrasonic echo signal, the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal, and the fourth feature of the ultrasonic echo signal have a unified dimension, so as to avoid the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal, and the fourth feature of the ultrasonic echo signal being inconsistent, resulting in some features dominating or some features being ignored, so as to provide a guarantee for reconstructing a high-resolution ultrasonic image and improve the accuracy of reconstructing a high-resolution ultrasonic image.

[0039] When the diagnostic reasoning unit matches the lesion feature with the disease feature knowledge base and calculates the matching degree, it includes: According to the pathological features, semantic similarity is calculated in the disease feature knowledge base according to the disease features to obtain feature similarity; Screening is performed based on the feature similarity combined with the first threshold, and disease features corresponding to feature similarities greater than the first threshold are screened out to obtain a first screening result; In the first screening results, the target disease is determined based on the disease characteristics; Determine the number of pathological features; When the number of pathological features is one, the disease features of the target disease are acquired to obtain the disease features of the target disease, and the disease influence analysis is performed on the disease features of the target disease to determine the influencing factors of the disease features, and the influencing factors of the disease features are combined with the feature similarity to obtain the disease matching degree; When the number of pathological features is two or more, the disease matching degree is calculated for the pathological features in combination with the target disease, and at the same time, the pathological features are analyzed to see whether there is any association, and the association analysis results are obtained; Combining the pathological features according to the association analysis results to determine the pathological feature combination results; The disease characteristics of the target disease are acquired to obtain the disease characteristics of the target disease, and the disease influence analysis is performed on the disease characteristics of the target disease to determine the influencing factors of the disease characteristics. At the same time, the pathological feature weight analysis is performed on the pathological feature combination results to determine the weight of the pathological feature. Then, the influencing factors of the disease characteristics are combined with the weight of the pathological characteristics and the feature similarity to obtain the disease matching degree.

[0040] Among them, when calculating the disease matching degree for the pathological features combined with the target disease in turn, the analysis and calculation method is the same as when the number of pathological features is one.

[0041] When pathological features are combined according to the association analysis results, if the association analysis results show that there is no association between the pathological features, there is no need to combine the pathological features. At this time, there is no pathological feature combination, and there is no need to perform subsequent analysis and calculation based on the pathological feature combination results.

[0042] When the influencing factors of disease characteristics are combined with the characteristic similarity to obtain the disease matching degree, it is calculated by the following formula:

[0043] In the above formula, Represents the disease feature knowledge base The matching degree between the disease and pathological characteristics, Indicates The first The influencing factors of disease characteristics, Pathological features and The first The feature similarity between the disease features.

[0044] The influencing factors of disease characteristics are combined with the weights of pathological characteristics and feature similarity to obtain the disease matching degree through the following formula:

[0045] In the above technical solution, Represents the disease feature knowledge base Disease and The matching degree of the pathological feature combination, Indicates The first The influencing factors of disease characteristics, Indicates The pathological feature combination The weight of each pathological feature, Indicates The pathological feature combination The pathological features and The first The feature similarity between disease features.

[0046] In the above embodiment, preliminary screening is achieved based on feature similarity through the first threshold, so that in the subsequent matching analysis and calculation process, only the corresponding diseases in the first screening results are analyzed and calculated, which reduces the workload of analysis and calculation and improves the efficiency of matching calculation acquisition. In addition, by determining the number of pathological features, different analyses are performed on different numbers of pathological features, thereby improving the comprehensiveness of the analysis and calculation. When the number of pathological features is two or more, in addition to calculating the disease matching degree for the pathological features combined with the target disease in turn, the pathological features are also analyzed to see whether there is a correlation, so that the pathological features are analyzed as a whole according to the correlation analysis results. The relationship between the pathological features is fully considered, the comprehensiveness of the diagnostic reasoning unit is improved, and then a guarantee is provided for generating a diagnostic conclusion, thereby improving the accuracy of the diagnostic conclusion. In addition, when determining the disease matching degree, the disease influence analysis is performed on the disease characteristics of the target disease, the different effects of different disease characteristics on the disease are considered, and the influencing factors of the disease characteristics are used to reconcile them in the process of disease matching calculation to improve the accuracy of disease matching. At the same time, in the process of calculating the disease matching degree by combining pathological characteristics, the relative importance of different pathological characteristics is also considered, and the weights of the pathological characteristics are used for adjustment, so that the disease matching degree is more objective and the accuracy of the disease matching is improved, thereby more accurately determining the disease type in the lesion area.

[0047] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A transesophageal visualization intelligent ultrasound probe, comprising an ultrasound probe (1) and a main support (2), characterized in that: The main support (2) is provided with a slide rail (3), a support plate (4) is slidably provided on the slide rail (3), probe handle mounting sleeves (12) are provided on both sides of the support plate (4), the ultrasound probe (1) can be snap-fitted into the probe handle mounting sleeve (12), an imaging system for manipulation is arranged in the ultrasound probe (1), the imaging system is arranged to reconstruct ultrasound images and generate comprehensive analysis results and diagnostic conclusions, and when generating the diagnostic conclusions, the disease matching degree is obtained by combining the influencing factors of the disease characteristics with the weights of the pathological characteristics and the feature similarity, so as to determine the disease type of the lesion area.

2. The transesophageal visualization intelligent ultrasound probe according to claim 1, characterized in that: The probe handle mounting sleeve (12) is connected to side pull rods (16) on both sides, a pulling shaft (17) is vertically arranged on the top of the side pull rod (16), and the top axle pin of the pulling shaft (17) is connected to a linkage shaft (19), two clamping plates (20) are arranged above the probe handle mounting sleeve (12), the middle part of the outer side of the clamping plate (20) is connected to the linkage shaft (19) through an axle pin, a guide strip (22) is arranged on the outer side of the clamping plate (20), guide plates (23) are arranged on both sides of the probe handle mounting sleeve (12), the guide strip (22) slides through the guide plate (23), and a return spring (18) is arranged at the bottom of the side pull rod (16).

3. The transesophageal visualization intelligent ultrasound probe according to claim 1, characterized in that: The slide rail (3) is provided with a slide groove (9), a slide block (10) is slidably arranged in the slide groove (9), a locking switch (13) is arranged on the slide block (10), the bottom surface of the support plate (4) is connected to the top of the slide block (10) through an axle pin, a support foot (14) is arranged at the bottom of the slide block (10), a groove (11) matching the support foot (14) is arranged on the bottom surface of the support plate (4), a folding frame (5) is arranged on one side of the main frame (2), a connecting plate (6) is arranged on the side of the folding frame (5) away from the main frame (2), a first chuck (7) and a second chuck (8) are arranged on the connecting plate (6), and the ultrasonic probe (1) can be clamped in the probe handle mounting sleeve (12), the first chuck (7) or the second chuck (8).

4. The transesophageal visualization intelligent ultrasound probe according to claim 1, characterized in that: The imaging system includes: an ultrasound signal acquisition module, an intelligent image reconstruction module, a multimodal fusion analysis module, a disease intelligent diagnosis module and an image optimization and annotation module; The ultrasonic signal acquisition module is configured to acquire ultrasonic echo signals through an ultrasonic probe (1) passing through the esophagus and convert the signals into digital signals; The intelligent image reconstruction module is configured to extract features of the ultrasonic echo signal after obtaining the ultrasonic echo signal, and reconstruct a high-resolution ultrasonic image; The multimodal fusion analysis module is configured to perform fusion analysis on the reconstructed ultrasound image and the patient's physiological parameter data to generate a comprehensive analysis result; The disease intelligent diagnosis module is configured to identify and judge the lesion characteristics in the ultrasound image based on the comprehensive analysis results and in combination with the disease characteristic knowledge base, and output a diagnosis conclusion. The disease characteristic knowledge base includes morphological characteristics, echo characteristics, hemodynamic parameters of different diseases on ultrasound images, and corresponding disease diagnosis basis, differential diagnosis information and treatment suggestions; The image optimization and annotation module is configured to perform targeted optimization processing on the ultrasound image according to the diagnosis conclusion, enhance the display effect of the lesion area, add annotation information related to the diagnosis result, and generate an annotated image.

5. The transesophageal visualization intelligent ultrasound probe according to claim 4, characterized in that: The multimodal fusion analysis module comprises: A feature matching unit is configured to extract image features layer by layer from the reconstructed ultrasound image, wherein the image features include edge contours, grayscale distribution, and texture features, model the patient's physiological parameter data using a time series and extract physiological parameter time series features, wherein the physiological parameter time series features include trend changes, extreme points, and periodicity, match the image features with the physiological parameter time series features, calculate a correlation score, and screen a feature combination; A data association unit, configured to convert image features and physiological parameter time series features into high-dimensional feature vectors, construct a graph structure between lesion features, image features and physiological parameter time series features, and define association weights of nodes and edges of the graph structure; The fusion calculation unit is configured to calculate a final fusion feature vector based on correlation scores of different modal features, mark abnormal data based on the fusion feature vector, and generate a comprehensive analysis result.

6. The transesophageal visualization intelligent ultrasound probe according to claim 4, characterized in that: The disease intelligent diagnosis module comprises: A lesion recognition unit is configured to perform pixel-level segmentation on the ultrasound image, extract possible lesion areas, perform secondary feature extraction on the lesion areas, obtain lesion features, wherein the lesion features include boundary morphology, signal intensity, and internal texture, and classify the lesion features into lesion types using a deep neural network; The diagnostic reasoning unit is configured to match the lesion features with the disease feature knowledge base and calculate the matching degree, determine the disease type of the lesion area, perform comprehensive reasoning based on rule reasoning combined with multiple lesion features, and generate a diagnostic conclusion.

7. The transesophageal visualization intelligent ultrasound probe according to claim 4, characterized in that: The image optimization and annotation module comprises: an adaptive enhancement unit configured to use histogram equalization to improve the visibility of image details in low-contrast lesion areas, enhance texture information in the lesion areas, and perform edge enhancement on the lesion areas of the ultrasound image; The intelligent labeling unit is configured to draw the lesion outline on the ultrasound image and label the corresponding disease type according to the segmentation result of the lesion area, and superimpose text information related to the disease type on the ultrasound image, wherein the text information includes the lesion size, type and severity, and generate a labeled image.

8. The transesophageal visualization intelligent ultrasound probe according to claim 5, characterized in that: The patient's physiological parameter data includes heart rate, blood pressure, blood oxygen saturation and respiratory rate, and the patient's physiological parameter data is pre-collected before the ultrasound image is collected.

9. The transesophageal visualization intelligent ultrasound probe according to claim 4, characterized in that: When the intelligent image reconstruction module extracts features from the ultrasonic echo signal, it includes: Performing signal preprocessing on the ultrasonic echo signal to obtain a first processed ultrasonic echo signal; Verify the first processed ultrasonic echo signal according to the time point, analyze whether the signals at the same time point are aligned, and obtain the verification analysis result; Performing alignment adjustment on the first processed ultrasonic echo signal according to the verification analysis result to obtain a second processed ultrasonic echo signal; Performing time domain feature analysis based on the second processed ultrasonic echo signal to obtain the time domain feature of the second processed ultrasonic echo signal and obtain a first feature of the ultrasonic echo signal; Converting the second processed ultrasonic echo signal from the time domain to the frequency domain, and performing frequency domain feature analysis to obtain a second feature of the ultrasonic echo signal; Decomposing the second processed ultrasonic echo signal, decomposing the second processed ultrasonic echo signal into a plurality of wavelet signals of different scales, and performing feature analysis on the wavelet signal to obtain a third feature of the ultrasonic echo signal; parsing the second processed ultrasonic echo signal, and analyzing the intensity change information of the second processed ultrasonic echo signal to obtain a fourth feature of the ultrasonic echo signal; Normalization analysis is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal, and normalization adjustment is performed on the first feature of the ultrasonic echo signal, the second feature of the ultrasonic echo signal, the third feature of the ultrasonic echo signal and the fourth feature of the ultrasonic echo signal to obtain the ultrasonic echo signal feature extraction result.

10. The transesophageal visualization intelligent ultrasound probe according to claim 6, characterized in that: When the diagnostic reasoning unit matches the lesion feature with the disease feature knowledge base and calculates the matching degree, it includes: According to the pathological features, semantic similarity is calculated in the disease feature knowledge base according to the disease features to obtain feature similarity; Screening is performed based on the feature similarity combined with the first threshold, and disease features corresponding to feature similarities greater than the first threshold are screened out to obtain a first screening result; In the first screening results, the target disease is determined based on the disease characteristics; Determine the number of pathological features; When the number of pathological features is one, the disease features of the target disease are acquired to obtain the disease features of the target disease, and the disease influence analysis is performed on the disease features of the target disease to determine the influencing factors of the disease features, and the influencing factors of the disease features are combined with the feature similarity to obtain the disease matching degree; When the number of pathological features is two or more, the disease matching degree is calculated for the pathological features in combination with the target disease, and at the same time, the pathological features are analyzed to see whether there is any association, and the association analysis results are obtained; Combining the pathological features according to the association analysis results to determine the pathological feature combination results; The disease characteristics of the target disease are acquired to obtain the disease characteristics of the target disease, and the disease influence analysis is performed on the disease characteristics of the target disease to determine the influencing factors of the disease characteristics. At the same time, the pathological feature weight analysis is performed on the pathological feature combination results to determine the weight of the pathological feature. Then, the influencing factors of the disease characteristics are combined with the weight of the pathological characteristics and the feature similarity to obtain the disease matching degree.

Images