A pre-hospital care assistance system based on wearable ultrasound equipment
By combining wearable ultrasound equipment and edge computing platforms with deep learning algorithms, the problem of difficulty in determining the cause of illness in pre-hospital emergency care has been solved, enabling rapid and accurate assessment of patient health status and disease risk, and supporting professional emergency care decision-making.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV QILU HOSPITAL
- Filing Date
- 2023-07-14
- Publication Date
- 2026-06-02
AI Technical Summary
In pre-hospital emergency care, it is difficult to determine the cause of a patient's illness, and there is a shortage of professional personnel, which leads to false alarms or delays in treatment.
Wearable ultrasound equipment, combined with edge computing platforms and deep learning algorithms, generates organ images through ultrasound phased array and transmits them to the hospital's ultrasound clinic, where deep learning networks are used for disease assessment.
It provides doctors with accurate assessments of patients' physical condition and disease risks in a short period of time, ensuring timely and effective emergency measures.
Smart Images
Figure CN117064433B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of pre-hospital emergency care, and in particular to a pre-hospital emergency care assistance system based on wearable ultrasound equipment. Background Technology
[0002] Pre-hospital emergency care refers to emergency care for critically ill patients outside of hospitals. In a broad sense, pre-hospital emergency care refers to emergency rescue carried out on-site by medical staff or bystanders when a patient falls ill. In a narrow sense, pre-hospital emergency care refers to the on-site rescue and en route monitoring carried out by professional emergency medical organizations equipped with communication equipment, transportation vehicles, and basic medical elements before the patient arrives at the hospital.
[0003] Pre-hospital emergency care is characterized by randomness, high mobility, poor emergency environment conditions, and a diverse and complex range of illnesses. Pre-hospital emergency care sometimes takes place in the wild, on the roadside, inside modified vehicles, or during transport, where light, noise, and vibration can make procedures such as auscultation and measurement of vital signs difficult. Because of the diverse and complex nature of the illnesses, and the need for paramedics to make a preliminary diagnosis and provide emergency treatment within a short timeframe, bystanders outside of hospitals may lack the necessary expertise to accurately and quickly assess the illness. This can lead to situations where the cause of the illness is misreported or the cause is unknown when calling 120, delaying the patient's treatment. Summary of the Invention
[0004] In order to make relatively accurate disease diagnosis in a short time, the present invention provides a pre-hospital emergency care assistance system based on wearable ultrasound equipment.
[0005] Firstly, the present invention provides a novel wearable ultrasound device under a mobile medical architecture, employing the following technical solution:
[0006] A novel wearable ultrasound device under a mobile medical architecture includes a wearable body and an ultrasound phased array, wherein the ultrasound phased array is coupled to the wearable body.
[0007] The ultrasonic phased array is electrically connected to an edge computing platform. The raw ultrasonic echo radio frequency data emitted from the ultrasonic phased array is transferred to the edge computing platform, which is configured to receive the raw ultrasonic echo radio frequency data and generate and display images of the monitored organs.
[0008] Furthermore, the edge computing platform is connected to the hospital's ultrasound examination room via 5G, and the edge computing platform is used to transmit images of the monitored organs to the hospital's ultrasound examination room.
[0009] Furthermore, the wearable body is made of a flexible and stretchable material, and the ultrasonic phased array is provided with multiple array elements, with the spacing between the multiple array elements set between 100um and 800um.
[0010] Furthermore, the wearable body is provided with multiple gel suction claws, which are used to adhere to the human body.
[0011] Secondly, the present invention provides a pre-hospital emergency care assistance system based on wearable ultrasound equipment. The edge computing platform is equipped with a deep learning ultrasound imaging algorithm, the input of which is raw ultrasound echo radio frequency data, and the output is an image of the monitored organ generated by the DNN.
[0012] Furthermore, the deep learning ultrasound imaging algorithm is as follows:
[0013] The original ultrasound echo radio frequency data is quantized according to the scan lines, number of array elements, and lateral sampling points. After quantization, knowledge distillation is performed. The knowledge-distilled data is copied and stored. One copy is compensated and summed to obtain the DAS ultrasound planar image. The other copy is based on the encoding-decoding model and combined with the DAS ultrasound planar image to obtain the DNN ultrasound image.
[0014] Furthermore, the method for acquiring the ultrasound planar image is as follows:
[0015] The knowledge distilled data is subjected to ToF compensation, followed by SUM, to obtain compensated RF data. The compensated RF data is then summed according to the array elements to form DAS data. The DAS data is then enveloped to form DAS ultrasound planar images in the scanning line and lateral sampling point directions.
[0016] Furthermore, the method for acquiring the DNN ultrasound planar image is as follows:
[0017] The knowledge distilled data is standardized to obtain RF data. The RF data is then processed using an Encoder-Decoder-like model and combined with DAS ultrasound planar images to form DNN ultrasound images in the scanning line and lateral sampling point directions.
[0018] Furthermore, the edge computing platform is equipped with a computer-aided diagnostic algorithm, which is as follows:
[0019] Quantitative analysis of DNN ultrasound images is performed. Feature extraction of DNN ultrasound images is carried out through joint space-time attention, which is divided into segmentation branch and classification branch. Aggregation loss is applied to the segmentation branch and the classification branch respectively, while periodic sequence information in the time dimension is introduced.
[0020] Furthermore, DNN ultrasound images are combined with deep learning network algorithms to obtain health status detection results and disease risk assessments.
[0021] In summary, the present invention has the following beneficial technical effects:
[0022] The wearable device attaches to the patient's body using gel suction claws. The raw ultrasound echo radio frequency data emitted by the ultrasound phased array on the wearable device is transmitted to the edge computing platform. The edge computing platform generates an image of the organ being examined and transmits it to the hospital's ultrasound clinic. After an emergency call is made, the doctor can quickly and accurately understand the patient's condition and respond promptly.
[0023] This invention quantizes raw ultrasound echo radio frequency data to generate a DAS (Digital Angiography) planar image. Based on the quantized parameters and the DAS planar image, a DNN (Digital Neural Network) ultrasound image is formed, replacing the traditional ultrasound imaging beam focusing method. This imaging method results in clearer images and more accurate detection results. Furthermore, after quantizing and analyzing the DNN ultrasound image, periodic sequence information is introduced to compensate for the limitations of ultrasound imaging quality in hospital ultrasound departments. Finally, a deep learning network algorithm is used to obtain health status detection results and disease risk assessment results, enabling doctors to make faster and more accurate disease diagnoses. Attached Figure Description
[0024] Figure 1 This is an overall structural block diagram of the present invention.
[0025] Figure 2 This is a block diagram of the deep learning ultrasound imaging algorithm of the present invention.
[0026] Figure 3 This is a flowchart of the computer-aided diagnostic algorithm of the present invention. Detailed Implementation
[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0028] Example 1
[0029] This invention provides a novel wearable ultrasound device within a mobile medical architecture, employing the following technical solution:
[0030] It includes a wearable body and an ultrasonic phased array, wherein the ultrasonic phased array is coupled to the wearable body.
[0031] The wearable device is similar to a semi-finished garment that can be worn on the human body. Made of flexible and stretchable material, unlike large equipment, it can be bent and stored away for easy storage in a car or for carrying around, and can be used in case of emergencies. Since people who have suddenly fallen or had an accident cannot be moved without professional guidance, the wearable device is equipped with multiple gel suction claws that directly adhere to the skin.
[0032] An ultrasound phased array consists of multiple array elements, with the spacing between these elements ranging from 100µm to 800µm. To maximize the ultrasound quality of the phased array, the element spacing should be similar to the length of the ultrasound wave. For example, for an ultrasound monitoring frequency of 2MHz for the heart, the element spacing would be approximately 500µm.
[0033] Reference Figure 1 An edge computing platform is electrically connected to the ultrasonic phased array. The raw ultrasonic echo radio frequency data emitted by the ultrasonic phased array is transferred to the edge computing platform. The raw ultrasonic echo radio frequency data is transmitted using a 5G network to the edge computing platform. The edge computing platform is used to receive the raw ultrasonic echo radio frequency data and generate and display images of the monitored organs. At the same time, it performs health status detection and disease risk assessment on the images of the monitored organs.
[0034] The edge computing platform connects to the hospital's ultrasound clinic via 5G. The edge computing platform is used to transmit images of the monitored organs, as well as health status detection and disease risk assessment results to the hospital's ultrasound clinic.
[0035] After the patient collapses, personnel equipped with the wearable ultrasound device attach it to the patient's body using a gel suction gripper. Simultaneously, they call 120 (emergency services). The ultrasound phased array focuses ultrasound on key points of the body's organs and transmits the raw ultrasound echo radio frequency data to an edge computing platform. The edge computing platform generates images of the monitored organs and analyzes their health status, transmitting all of this information to the hospital's ultrasound clinic. Upon receiving the call, medical staff retrieve the transmitted images and analysis results, responding promptly and accurately from a professional medical perspective to provide the patient with the optimal rescue plan.
[0036] Example 2
[0037] This invention provides a pre-hospital emergency care assistance system based on wearable ultrasound equipment, as detailed below:
[0038] The edge computing platform is equipped with a deep learning ultrasound imaging algorithm. The input is raw ultrasound echo radio frequency data, and the output is an image of the monitored organ generated by the DNN.
[0039] Reference Figure 2 The deep learning ultrasound imaging algorithm is as follows:
[0040] The raw ultrasound echo radio frequency data is quantized according to the scan lines, number of array elements and lateral sampling points. After quantization, knowledge distillation is performed to reduce the quantization of model parameters. The knowledge-distilled data is then copied and stored.
[0041] One of the data is subjected to ToF compensation, and then SUM is performed to obtain compensated RF data. The compensated RF data is summed according to the array elements to form DAS data. The DAS data is then enveloped to form DAS ultrasound planar images in the scanning line and lateral sampling point direction.
[0042] Another set of data is standardized to obtain RF data. The RF data is then processed using an Encoder-Decoder-like model and combined with DAS ultrasound planar images to form DNN ultrasound images in the direction of scanning lines and lateral sampling points.
[0043] Reference Figure 3 The edge computing platform is equipped with a computer-aided diagnostic algorithm, which is as follows:
[0044] Quantitative analysis of DNN ultrasound images is performed, and features are extracted from the DNN ultrasound images through joint space-time attention, which is divided into segmentation branches and classification branches. Aggregation loss is applied to the segmentation branches and classification branches respectively, and periodic sequence information in the time dimension is introduced, which can make up for the problem of ultrasound imaging quality in hospital ultrasound departments.
[0045] Because 5G transmission may result in distortion or misalignment of image segments, leading to inaccuracies in ultrasound imaging in hospital departments, DNN ultrasound images are segmented into branches, and time-dimensional periodic sequence information is introduced. This is equivalent to labeling the image segments with time-dimensional information. After transmission, the segments are aggregated according to the dimensional information to compensate for the quality issues of ultrasound imaging in hospital departments.
[0046] Finally, a deep learning network algorithm is used to obtain health status detection results and disease risk assessment.
[0047] The implementation principle of a pre-hospital emergency care system based on wearable ultrasound equipment according to an embodiment of the present invention is as follows: After a patient falls to the ground, a person with the wearable ultrasound equipment attaches the equipment to the patient's body using a gel suction claw, and simultaneously dials 120 for emergency medical assistance. The ultrasound phased array begins detection and transmits the raw ultrasound echo radio frequency data to an edge computing platform. The edge computing platform obtains images of the monitored organs and health status analysis results, and transmits them all to the hospital's ultrasound clinic. Upon receiving the call, medical staff can retrieve the images and analysis results transmitted at that time and react promptly and accurately from the perspective of professional medical staff, providing the patient with the optimal rescue plan.
[0048] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A wearable ultrasound device under a novel mobile medical architecture, characterized in that: It includes a wearable body and an ultrasonic phased array, wherein the ultrasonic phased array is coupled to the wearable body; The ultrasonic phased array is electrically connected to an edge computing platform. The raw ultrasonic echo radio frequency data emitted from the ultrasonic phased array is transferred to the edge computing platform, which is configured to receive the raw ultrasonic echo radio frequency data and generate and display images of the monitored organs. The edge computing platform is equipped with a deep learning ultrasound imaging algorithm. The input is raw ultrasound echo radio frequency data, and the output is an image of the monitored organ generated by the DNN. The deep learning ultrasound imaging algorithm is as follows: the original ultrasound echo radio frequency data is quantized according to the scanning rows, the number of array elements and the horizontal sampling points. After quantization, knowledge distillation is performed. The knowledge-distilled data is copied and stored. One copy is compensated and then summed to obtain the DAS ultrasound planar image. The other copy is based on the encoding-decoding model and combined with the DAS ultrasound planar image to obtain the DNN ultrasound image. The method for acquiring the ultrasound planar image is as follows: the knowledge distilled data is subjected to ToF compensation, then SUM is performed, and the compensated RF data is obtained. The compensated RF data is summed according to the array elements to form DAS data. The DAS data is enveloped to form DAS ultrasound planar images in the scanning line and lateral sampling point direction. The method for acquiring DNN ultrasound images is as follows: the data after knowledge distillation is standardized to obtain RF data, and the RF data is combined with DAS ultrasound planar images through an Encoder-Decoder-like model to form DNN ultrasound images in the direction of scanning lines and lateral sampling points. The computer-aided diagnostic algorithm is as follows: quantitative analysis of DNN ultrasound images, feature extraction of DNN ultrasound images through joint space-time attention, divided into segmentation branches and classification branches, aggregation loss is applied to segmentation branches and classification branches respectively, and periodic sequence information in the time dimension is introduced.
2. The wearable ultrasound device under a novel mobile medical architecture according to claim 1, characterized in that: The edge computing platform is connected to the hospital's ultrasound clinic via 5G, and is used to transmit images of the monitored organs to the hospital's ultrasound clinic.
3. The wearable ultrasound device under a novel mobile medical architecture according to claim 2, characterized in that: The wearable body is made of a flexible and stretchable material, and the ultrasonic phased array is provided with multiple array elements, with the spacing between the multiple array elements set between 100um and 800um.
4. The wearable ultrasound device under a novel mobile medical architecture according to claim 1, characterized in that: The wearable body is equipped with multiple gel suction claws, which are used to adhere to the human body.
5. A pre-hospital emergency care assistance system based on wearable ultrasound equipment according to claim 4, characterized in that: DNN ultrasound images combined with deep learning network algorithms yield health status detection results and disease risk assessments.