Intelligent hemostasis system based on VR technology

Abstract

The invention discloses an intelligent hemostasis system based on the VR technology, and relates to the field of VR visualization, and the system comprises a physiological monitoring module which collects the blood vessel pressure, geometric morphology and blood flow data of a target blood vessel; the AI decision module determines a blood vessel state by utilizing a blood vessel state prediction model according to the collected data, determines a bleeding point condition according to a bleeding point 3D model, and selects a hemostasis mode according to the bleeding point condition; the VR man-machine interaction module comprises a 3D imaging unit, a VR interaction interface and a controller; the 3D imaging unit scans a bleeding point in real time by using a micro endoscope and a depth sensor, and constructs a bleeding point 3D model; the 3D model of the bleeding point is displayed on the VR interaction interface, and a doctor operates the virtual hemostasis tool through the controller so as to control the intelligent hemostasis actuator to stop bleeding in the hemostasis mode. According to the application, high-precision, self-adaptive and visual hemostasis operation is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of VR visualization, in particular to an intelligent hemostasis system based on VR technology. BACKGROUND

[0002] Traditional hemostasis methods (such as electrocoagulation, compression, suturing, etc.) rely on the experience of doctors, and in complex operations or emergency situations, there may be problems of inaccurate hemostasis, damage to surrounding tissues, or low operation efficiency. VR technology has been applied to surgical simulation and remote operation, but has not been effectively integrated into real-time hemostasis control. SUMMARY

[0003] The purpose of the present application is to provide an intelligent hemostasis system based on VR technology to realize high-precision, adaptive, and visual hemostasis operation.

[0004] To achieve the above purpose, the present application provides the following solutions: In a first aspect, the present application provides an intelligent hemostasis system based on VR technology, comprising a physiological monitoring module, an AI decision module, a VR human-computer interaction module, and an intelligent hemostasis executor. The physiological monitoring module is used to collect blood vessel pressure, geometric morphology, and blood flow data of a target blood vessel; the blood flow data includes blood flow velocity, blood temperature, and tissue oxygenation state. The AI decision module is used to determine blood vessel state by using a blood vessel state prediction model according to the blood vessel pressure, the geometric morphology, and the blood flow data, and is also used to determine bleeding point condition according to a bleeding point 3D model and select a hemostasis mode according to the bleeding point condition; the blood vessel state is normal, hardened, or neoplastic; the bleeding point condition is venous bleeding, arterial bleeding, or capillary bleeding; the hemostasis mode is electrocoagulation mode, ultrasonic hemostasis mode, or biological glue injection mode. The VR human-computer interaction module includes a 3D imaging unit, a VR interaction interface, and a controller; the 3D imaging unit is used to scan a bleeding point in real time by using a miniature endoscope and a depth sensor and construct the bleeding point 3D model; the VR interaction interface is used to display the bleeding point 3D model, and a doctor operates a virtual hemostasis tool by the controller to control the intelligent hemostasis executor to perform hemostasis by using the hemostasis mode.

[0005] Optionally, the AI decision module includes: a blood vessel state prediction unit, configured to determine blood vessel state by using a blood vessel state prediction model according to the blood vessel pressure, the geometric morphology, and the blood flow data; a bleeding point condition determination unit, configured to determine bleeding point condition according to a bleeding point 3D model; a hemostasis mode selection unit, configured to select a hemostasis mode according to bleeding point condition.

[0006] Optionally, the vascular state prediction unit includes: A data acquisition subunit is used to acquire the vascular pressure, the geometry, and the blood flow data; The stress distribution determination subunit is used to establish a vascular biomechanical model based on the vascular pressure, geometry and blood flow data, and to calculate the stress distribution of the target vascular vessel. The feature extraction subunit is used to extract pressure time-frequency features and image features based on the stress distribution of the target blood vessel; The vascular condition prediction subunit is used to determine the vascular condition and rupture risk score based on the pressure time-frequency characteristics and image characteristics, using a vascular condition prediction model combined with a physical constraint optimization model; wherein, the vascular condition prediction model is obtained by training a dual-channel neural network using a training dataset.

[0007] Optionally, the physiological monitoring module includes a pressure sensor; The pressure sensor is used to collect the vascular pressure of the target blood vessel.

[0008] Optionally, the physiological monitoring module further includes: a velocity sensor, a temperature sensor, and a tissue oxygenation sensor; The velocity sensor is used to collect blood flow velocity; The temperature sensor is used to collect blood temperature; The tissue oxygenation sensor is used to detect tissue oxygenation status.

[0009] Optionally, the AI ​​decision-making module further includes: The bleeding risk prediction unit is used to assess whether there is a bleeding risk in the target blood vessel based on the blood flow velocity, the blood temperature, and the tissue oxygenation status.

[0010] Optionally, the intelligent hemostasis actuator includes a robotic arm and a hemostasis head; The robotic arm is controlled by the doctor's gestures or the controller, thereby controlling the hemostatic head.

[0011] Optionally, the intelligent hemostasis actuator further includes: a pressure feedback unit; The pressure feedback unit is connected to the robotic arm and the physiological monitoring module respectively; the pressure feedback unit is used to adjust the pressure of the robotic arm according to the pressure of the hemostasis point.

[0012] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides an intelligent hemostasis system based on VR technology, including a physiological monitoring module, an AI decision-making module, a VR human-computer interaction module, and an intelligent hemostasis actuator. The physiological monitoring module collects vascular pressure, geometry, and blood flow data of the target blood vessel. The AI ​​decision-making module determines the vascular state based on the vascular pressure, geometry, and blood flow data using a vascular state prediction model, and also determines the bleeding point condition based on a 3D model of the bleeding point, selecting a hemostasis mode accordingly. The VR human-computer interaction module includes a 3D imaging unit, a VR interface, and a controller. The 3D imaging unit uses a miniature endoscope and depth sensor to scan the bleeding point in real time and construct a 3D model of the bleeding point. The VR interface displays the 3D model of the bleeding point, and the doctor operates the virtual hemostasis tool through the controller to control the intelligent hemostasis actuator to perform hemostasis using the described hemostasis mode. This VR-based intelligent hemostasis system achieves high-precision, adaptive, and visualized hemostasis operations. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of an intelligent hemostasis system based on VR technology, provided as an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] The intelligent hemostasis system based on VR technology provided in this application is a comprehensive system that combines virtual reality visualization, real-time biofeedback, and intelligent hemostasis control.

[0018] In one exemplary embodiment, such as Figure 1 As shown, a VR-based intelligent hemostasis system is provided, including: a physiological monitoring module, an AI decision-making module, a VR human-computer interaction module, and an intelligent hemostasis actuator.

[0019] The physiological monitoring module is used to acquire vascular pressure, geometry, and blood flow data of the target blood vessel. The blood flow data includes blood flow velocity, blood temperature, and tissue oxygenation status. Imaging techniques for acquiring vascular geometry include intravascular ultrasound (IVUS), optical coherence tomography (OCT), and angiography.

[0020] As an optional implementation, the physiological monitoring module includes a pressure sensor, a velocity sensor, a temperature sensor, and a tissue oxygenation sensor.

[0021] The pressure sensor is used to collect the vascular pressure of the target blood vessel. In practical applications, when the pressure sensor detects a change in vascular pressure, it can issue an early warning to the operator (doctor) indicating a change in the condition of the blood vessel; or when bleeding occurs during surgical procedures, the pressure sensor can be used to quickly locate the hemostasis point.

[0022] The velocity sensor is used to collect blood flow velocity.

[0023] The temperature sensor is used to collect blood temperature.

[0024] The tissue oxygenation sensor is used to detect tissue oxygenation status.

[0025] The AI ​​decision-making module is used to determine the vascular state based on the vascular pressure, the geometric shape, and the blood flow data using a vascular state prediction model. It is also used to determine the bleeding point condition based on the 3D model of the bleeding point and select a hemostasis mode based on the bleeding point condition. The vascular state is normal, sclerotic, or aneurysmal. The bleeding point condition is venous bleeding, arterial bleeding, or capillary bleeding. The hemostasis mode is electrocoagulation mode, ultrasonic hemostasis mode, or bio-adhesive spray mode.

[0026] As an optional implementation, the AI ​​decision-making module includes: The vascular state prediction unit is used to determine the vascular state based on the vascular pressure, the geometric shape, and the blood flow data using a vascular state prediction model.

[0027] The bleeding point condition determination unit is used to determine the bleeding point condition based on the 3D model of the bleeding point.

[0028] The hemostasis mode selection unit is used to select the hemostasis mode according to the condition of the bleeding point.

[0029] As an optional implementation, the vascular state prediction unit includes: The data acquisition subunit is used to acquire the vascular pressure, the geometry, and the blood flow data.

[0030] The stress distribution determination subunit is used to establish a vascular biomechanical model based on the vascular pressure, geometry and blood flow data, and to calculate the stress distribution of the target vascular vessel.

[0031] The feature extraction subunit is used to extract pressure time-frequency features and image features based on the stress distribution of the target blood vessel.

[0032] The specific process for extracting time-frequency and image features of pressure is as follows: The feature extraction subunit for constructing the vascular pressure prediction model needs to extract key features from pressure signals and image data and correlate them with vascular stress distribution. The entire process begins with the intravascular pressure waveform acquired by a high-precision pressure sensor, and extracts time-domain and frequency-domain features through signal processing techniques. In time-domain analysis, the peak values ​​of systolic and diastolic pressure waveforms are automatically identified, and parameters such as pulse pressure difference, pressure change rate, and enhancement index are calculated. These parameters directly reflect the elasticity of the blood vessel and the pulse wave reflection characteristics. Frequency-domain analysis performs Fourier transform on the pressure signal to decompose the fundamental frequency and harmonic components, and evaluates the impedance characteristics of the blood vessel by analyzing the energy distribution of each frequency band. In terms of image feature extraction, the system processes image data from intravascular ultrasound or optical coherence tomography. First, the vascular image is segmented to accurately identify the lumen boundary and vascular wall structure, and geometric parameters such as lumen area, vascular wall thickness, and its distribution uniformity are calculated accordingly. At the same time, texture analysis methods are used to quantify the texture characteristics of plaque regions through algorithms such as gray-level co-occurrence matrix, distinguishing the mechanical properties of different plaque components. These morphological features, combined with the vascular stress distribution results obtained through finite element analysis, extract mechanical parameters such as maximum equivalent stress and stress gradient, establishing a correlation model between vascular structure and mechanical response. Finally, the extracted time-frequency features, imaging features, and mechanical features are normalized and fused. Dimensionality reduction methods such as principal component analysis are used to integrate multidimensional features into more representative feature vectors. The entire process achieves automated conversion from raw data to feature vectors, providing highly informative input data for subsequent vascular state prediction. This multimodal feature extraction method not only preserves the specific information of various data types but also reveals the intrinsic relationship between vascular physiological state and mechanical behavior through coupling analysis between features.

[0033] The vascular condition prediction subunit is used to determine the vascular condition and rupture risk score based on the pressure time-frequency characteristics and image characteristics, using a vascular condition prediction model combined with a physical constraint optimization model; wherein, the vascular condition prediction model is obtained by training a dual-channel neural network using a training dataset.

[0034] In practical applications, pressure sensors are used to acquire pressure data of blood vessels in different states, generating raw pressure data and the correlation between pressure and blood vessel state. This allows for the construction of a model that predicts blood vessel state using blood pressure (blood vessel state prediction model). The specific method involves: first, acquiring blood vessel pressure waveforms, geometric morphology, and blood flow data using pressure sensors and imaging equipment; second, establishing a blood vessel biomechanical model and calculating stress distribution; then, extracting time-frequency features of pressure (such as pulse pressure difference and enhancement index) and imaging features (such as plaque volume), and inputting them into a dual-channel neural network (LSTM+3D CNN) for training; finally, incorporating finite element simulation results as a physical constraint optimization model, and outputting a blood vessel state classification (normal / sclerotic / tumorous) and a rupture risk score.

[0035] The finite element simulation results are obtained through biomechanical simulation based on patient-specific vascular geometry. The specific process is as follows: First, a three-dimensional vascular model is reconstructed using imaging data such as IVUS, OCT, or CTA (CT angiography), including details such as the lumen contour, layered structure of the vessel wall, and plaque morphology, and different material properties are assigned to different tissue regions (e.g., hyperelastic parameters of normal vessel walls and heterogeneous mechanical properties of plaques). Then, the blood pressure waveform measured by a pressure sensor is applied as a dynamic load to the vessel wall, considering blood flow shear force and surrounding tissue constraints. The finite element solver calculates key mechanical parameters such as the stress-strain distribution of the vessel wall, stress concentration at plaque boundaries, and fatigue damage accumulation under cyclic loading. These simulation results are integrated into the neural network training process as physical constraints, ensuring that the model predictions conform to mechanical laws through a biomechanical consistency loss function. Furthermore, features such as maximum principal stress and strain energy density are fused with imaging features, ultimately used to calibrate the vascular rupture risk score. The determination of high-risk areas requires both exceeding the threshold of mechanical parameters and meeting the conditions of imaging vulnerability characteristics. This method couples the patient’s actual anatomical structure, physiological load, and material properties, making the prediction results both data-driven sensitivity and interpretability based on physical laws.

[0036] Determining the risk score for vascular rupture is a multi-dimensional, multi-parameter fusion intelligent assessment process that quantitatively integrates biomechanical properties and clinicopathological features through a machine learning model. The specific implementation process is as follows: First, the time-frequency features of pressure (such as pulse pressure, enhancement index, and harmonic energy distribution) and imaging features (such as luminal stenosis rate, plaque vulnerability score, and stress concentration coefficient) obtained from the feature extraction subunit are input into a pre-trained dual-channel neural network. The spatiotemporal feature fusion layer of this network automatically calculates the weight association of each feature, with mechanically sensitive features (such as peak circumferential stress and stress gradient) receiving higher weights through an attention mechanism. Second, the model is combined with a physical constraint optimization module for biomechanical verification. Based on the stress-strain distribution of the blood vessel wall generated by finite element simulation, the system verifies the rationality of the mechanical parameters predicted by the neural network. If the deviation between the predicted stress concentration area and the actual mechanical simulation results exceeds 15%, an iterative optimization algorithm is triggered to readjust the feature weights, ensuring that the prediction results conform to the basic principles of solid mechanics. Ultimately, the calculation of the rupture risk score integrates three key indicators: (1) mechanical risk factors: including the ratio of maximum equivalent stress to the tensile strength of the vessel wall (stress risk coefficient) and the area ratio of stress concentration areas; (2) pathological risk factors: plaque vulnerability index calculated based on imaging features (including lipid core ratio, fibrous cap thickness, etc.); (3) dynamic risk factors: fatigue cumulative damage caused by pressure fluctuations (the influence of pressure cycle number and amplitude is quantified by rainflow counting method). The score output adopts a standardized scale of 0-100 points, where: 0-30 points correspond to low risk (annual rupture probability <5%); 31-70 points correspond to medium risk (annual probability 5-20%); and 71-100 points correspond to high risk (annual probability >20%).

[0037] In another embodiment, based on computer vision to identify bleeding points and the distribution of surrounding blood vessels, endoscopy, near-infrared imaging, laser speckle contrast imaging and depth sensors are used to enhance vascular contrast and detect real-time blood flow velocity and pressure. If vascular contrast, blood flow velocity and blood pressure exceed the normal range, there is a risk of bleeding.

[0038] As an optional implementation, the AI ​​decision-making module further includes: The bleeding risk prediction unit is used to assess whether there is a bleeding risk in the target blood vessel based on the blood flow velocity, the blood temperature, and the tissue oxygenation status.

[0039] Parameters such as blood flow velocity, tissue oxygenation, and blood temperature can be used to assess the risk of bleeding in blood vessels. When bleeding occurs, the values ​​of these parameters are outside the normal range.

[0040] The VR human-computer interaction module includes a 3D imaging unit, a VR interactive interface, and a controller. The 3D imaging unit is used to scan bleeding points in real time using a miniature endoscope and a depth sensor to construct a 3D model of the bleeding point. The VR interactive interface is used to display the 3D model of the bleeding point. Doctors operate virtual hemostasis tools through the controller to control the intelligent hemostasis actuator to perform hemostasis using the hemostasis mode.

[0041] In this embodiment, the VR human-computer interaction module also has the following functions: Gesture control: Doctors can adjust the position of hemostatic instruments using gestures.

[0042] Haptic feedback: When important tissues and blood vessels are detected in the field of view of the endoscope at the front end, the VR controller provides vibration feedback.

[0043] Voice commands: Supports switching hemostasis mode via voice commands.

[0044] In practical applications, after a bleeding point is detected, a miniature endoscope and a depth sensor are used to scan the wound (bleeding point) in real time and construct a 3D model of the bleeding point to predict its condition (venous bleeding, arterial bleeding, capillary bleeding (diffuse bleeding)). Based on the different bleeding point conditions, the hemostasis mode (electrocoagulation mode, ultrasonic hemostasis, and bio-glue spray) is intelligently selected.

[0045] Electrocoagulation mode: suitable for hemostasis in small blood vessels (veins).

[0046] Ultrasonic hemostasis: suitable for cutting larger blood vessels (arteries) or tissues.

[0047] Bio-adhesive spray: Suitable for diffuse bleeding (capillary bleeding).

[0048] Doctors wear VR headsets to observe magnified virtual surgical scenes and can operate virtual hemostasis tools through gestures or controllers to achieve rapid and precise hemostasis.

[0049] As an alternative implementation, the intelligent hemostasis actuator includes a robotic arm, a hemostasis head, and a pressure feedback unit.

[0050] The robotic arm is controlled by the doctor's gestures or the controller, thereby controlling the hemostatic head.

[0051] The pressure feedback unit is connected to the robotic arm and the physiological monitoring module respectively; the pressure feedback unit is used to adjust the pressure of the robotic arm according to the pressure of the hemostasis point.

[0052] In practical applications, pressure sensors monitor the force at the hemostasis point in real time and automatically adjust the pressure of the robotic arm to avoid excessive compression.

[0053] The intelligent hemostasis system based on VR technology described in this application can promptly alert doctors to high-risk situations during surgical procedures, prompting them to operate with caution and preventing bleeding (this is achieved through a physiological monitoring module and an AI decision-making module). In the event of bleeding, the system can quickly locate the bleeding point using pressure sensors and VR technology, and formulate a reasonable hemostasis plan based on the real-time bleeding status (achieved through a VR human-computer interaction module and an intelligent hemostasis actuator).

[0054] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0055] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A VR technology-based intelligent hemostasis system, characterized in that, The application relates to a physiological monitoring module, an AI decision module, a VR human-computer interaction module and an intelligent hemostasis executor. The physiological monitoring module is used for collecting blood vessel pressure, geometric shape and blood flow data of a target blood vessel; the blood flow data comprises blood flow velocity, blood temperature and tissue oxygenation state. The AI decision module is used for determining a blood vessel state by using a blood vessel state prediction model according to the blood vessel pressure, the geometric shape and the blood flow data, and is also used for determining a bleeding point condition according to a bleeding point 3D model and selecting a hemostasis mode according to the bleeding point condition; the blood vessel state is normal, hardened or neoplastic; the bleeding point condition is venous bleeding, arterial bleeding or capillary bleeding; and the hemostasis mode is an electrocoagulation mode, an ultrasonic hemostasis mode or a biological glue injection mode. The VR human-computer interaction module comprises a 3D imaging unit, a VR interaction interface and a controller; the 3D imaging unit is used for scanning a bleeding point in real time by using a miniature endoscope and a depth sensor and constructing the bleeding point 3D model; and the VR interaction interface is used for displaying the bleeding point 3D model, and a doctor operates a virtual hemostasis tool by using the controller to control the intelligent hemostasis executor to perform hemostasis by using the hemostasis mode. The AI decision module comprises:

2. The VR technology-based intelligent hemostasis system according to claim 1, wherein, a blood vessel state prediction unit used for determining a blood vessel state by using a blood vessel state prediction model according to the blood vessel pressure, the geometric shape and the blood flow data; a bleeding point condition determination unit used for determining a bleeding point condition according to a bleeding point 3D model; a hemostasis mode selection unit used for selecting a hemostasis mode according to a bleeding point condition. The blood vessel state prediction unit comprises:

3. The VR technology-based intelligent hemostasis system according to claim 2, wherein, a data acquisition subunit used for acquiring the blood vessel pressure, the geometric shape and the blood flow data; a stress distribution determination subunit used for establishing a blood vessel biomechanics model according to the blood vessel pressure, the geometric shape and the blood flow data and calculating stress distribution of the target blood vessel; a feature extraction subunit used for extracting pressure time-frequency features and image features according to the stress distribution of the target blood vessel; a blood vessel state prediction subunit used for determining a blood vessel state and a rupture risk score by using a blood vessel state prediction model in combination with a physical constraint optimization model according to the pressure time-frequency features and the image features; wherein the blood vessel state prediction model is obtained by training a double-channel neural network by using a training data set. The physiological monitoring module comprises a pressure sensor.

4. The VR technology based intelligent hemostatic system according to claim 1, wherein, The pressure sensor is used for collecting blood vessel pressure of the target blood vessel. The physiological monitoring module further comprises a velocity sensor, a temperature sensor and a tissue oxygenation sensor.

5. The VR technology based intelligent hemostatic system according to claim 1, wherein, The velocity sensor is used for collecting blood flow velocity. The temperature sensor is used for collecting blood temperature. The tissue oxygenation sensor is used for detecting tissue oxygenation state. The AI decision module further comprises:

6. The VR technology-based intelligent hemostatic system according to claim 4, wherein, a bleeding risk prediction unit used for evaluating whether the target blood vessel has a bleeding risk according to the blood flow velocity, the blood temperature and the tissue oxygenation state. The intelligent hemostasis executor comprises a mechanical arm and a hemostasis head.

7. The VR technology based intelligent hemostatic system according to claim 1, wherein, The mechanical arm is controlled by a doctor's gesture or the controller to control the hemostasis head. ​ 8. The VR technology-based intelligent hemostasis system according to claim 7, characterized in that, The intelligent hemostasis execution device further comprises a pressure feedback unit; The pressure feedback unit is connected with the mechanical arm and the physiological monitoring module respectively, and is used for adjusting the pressure of the mechanical arm according to the pressure of the hemostasis point.

Images