A method and system for automated assessment of focal ultrasound of trauma

Abstract

The application discloses a kind of automatic trauma focus ultrasonic evaluation method and system, wherein the system includes: data acquisition module, receive the ultrasonic echo signal of human tissue, the received ultrasonic echo signal is sampled by analog-digital converter and forms data stream;Signal processing module, data stream is demodulated to base frequency, and signal is handled by low-pass filter and logarithmic compression, and the signal after processing is obtained;Data transmission module, the signal after processing is transmitted to eFast auxiliary diagnosis module and ultrasonic imaging module;eFast auxiliary diagnosis module, the ultrasonic image corresponding to the signal after processing is divided into trace, small amount, medium amount and large amount according to the free degree of liquid, and the diagnostic result is obtained;Ultrasonic imaging module, scanning conversion, digital gain compensation, dynamic range adjustment and post-processing in image domain are carried out to ultrasonic image, and the processed ultrasonic image is obtained;Result output module, the diagnostic result and the processed ultrasonic image are fused, and the evaluation examination result is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of emergency medicine, and in particular to a method and system for automatic trauma focused ultrasound assessment. BACKGROUND

[0002] Focused assessment with sonography in trauma (FAST) is the most important work for emergency physicians to quickly assess the condition of patients with acute chest and abdominal closed injury at bedside. The traditional FAST examination mainly uses ultrasound to quickly judge whether there is free fluid in the abdominal cavity, and the extended eFAST examination content is extended to the detection of chest and pericardium. In trauma treatment, eFAST can quickly answer three main questions: 1. Is there pericardial effusion (blood); 2. Is there free fluid (blood) in the abdominal cavity; 3. Is there a hemothorax.

[0003] Normal thoracic and abdominal cavities contain a small amount of physiological fluid to facilitate the free sliding of organs in the thoracic and abdominal cavities. The patient's position, the source, nature, aggregation time of pathological fluid, and the degree of anatomical variation determine the position of free fluid in the abdominal cavity under ultrasound exploration. Free fluid is mostly aggregated in gravity-dependent areas, which appears as anechoic areas on ultrasound images.

[0004] Generally speaking, all free fluids on ultrasound images, including ascites, blood, bile, urine and lymph, are displayed as black. Blood clots and separated effusions, especially pus, often show stronger echoes due to higher protein content. The echoes of solid fragments after the rupture of hollow organs are uneven.

[0005] In 2017, the American College of Emergency Physicians (ACEP) summarized the advantages of eFAST in trauma treatment: ① Quickly identify life-threatening trauma such as pneumothorax, hemothorax or cardiac tamponade. ② Assist in deciding whether to perform other imaging examinations such as X-ray or CT. ③ Guide whether to go directly to the operating room or transfer to a higher level trauma center. ④ No radiation damage, can dynamically assess trauma patients [Chin J Emerg Med, January 2020, Vol. 29, No. 1].

[0006] eFast is a purposeful, rapid and effective ultrasound evaluation for critically ill patients, and the examination results can provide the basis for doctors to make or adjust clinical diagnosis and treatment decisions. However, in actual operation, due to the medical emergency of the patient, the patient's condition is critical and complex, and part of the patients are in a coma or passive position after trauma, which cannot effectively cooperate with the examination, so that the ultrasound evaluation of the patient by the clinician is often not comprehensive enough; and the judgment of the patient's condition by the clinician combined with clinical experience is usually subjective, and there will inevitably be deviations. Therefore, an objective, rapid and effective automatic eFast evaluation method is needed to seize the "golden period" of treatment of critically ill trauma patients, so as to maximize the mortality rate of trauma patients. SUMMARY

[0007] The purpose of the present application is to overcome the defects of the prior art, provide a method and system for automatic trauma key ultrasound evaluation, which can be well combined with existing ultrasound systems to meet the needs of clinicians to quickly and accurately evaluate acute chest and abdominal closed injuries and effectively carry out follow-up diagnosis and treatment.

[0008] In order to achieve the above purpose, the present application adopts the following technical scheme:

[0009] An automatic trauma key ultrasound evaluation system comprises:

[0010] A data acquisition module is used to receive ultrasound echo signals related to human tissues, sample the received ultrasound echo signals through an analog-to-digital converter and form a data stream, and transmit the formed data stream to a signal processing module;

[0011] A signal processing module is connected with the data acquisition module, used to receive the data stream, demodulate the received data stream to a base frequency, and process the signals corresponding to the data stream demodulated to the base frequency through a low-pass filter and a logarithmic compression to obtain processed signals;

[0012] A data transmission module is connected with the signal processing module, used to transmit the obtained processed signals to an eFast auxiliary diagnosis module and an ultrasound imaging module;

[0013] An eFast auxiliary diagnosis module is connected with the data transmission module, used to receive the processed signals, and use an artificial intelligence analysis algorithm of texture features of ultrasound images to divide the ultrasound images corresponding to the processed signals into trace, small amount, medium amount and large amount according to the degree of freedom of liquid to obtain a diagnosis result;

[0014] An ultrasound imaging module is connected with the data transmission module, used to receive the processed signals, and perform scan conversion, digital gain compensation, dynamic range adjustment and image domain post-processing on the ultrasound images corresponding to the processed signals to obtain processed ultrasound images;

[0015] An output module is connected with the eFast auxiliary diagnosis module and the ultrasonic imaging module, and is configured to fuse the diagnosis result obtained by the eFast auxiliary diagnosis module and the processed ultrasonic image obtained by the ultrasonic imaging module, so as to obtain an evaluation examination result corresponding to the ultrasonic image.

[0016] Further, the eFast auxiliary diagnosis module specifically comprises:

[0017] A segmentation module is configured to segment the ultrasonic image into a plurality of sub-blocks, and each sub-block is an independent training sample.

[0018] An initialization module is configured to initialize the weight distribution of the training sample, so as to obtain the weight D1 of each sample in the first iteration.

[0019] An iteration module is configured to perform multiple iterations to obtain a weak classifier, wherein the weight D is used in each iteration. m The training sample set is learned, and m represents the number of iterations.

[0020] A combination module is configured to combine the weak classifier into a strong classifier after the iteration is completed.

[0021] A classification module is configured to divide the ultrasonic image into trace, small amount, medium amount and large amount according to the degree of free liquid according to the strong classifier, so as to judge the severity of the damage.

[0022] Further, the initialization module is configured to initialize the weight distribution of the training sample, so as to obtain the weight of each sample in the first iteration.

[0023] The initialized weight w 1i is represented as:

[0024] ,

[0025] Wherein, N represents the image of the training sample.

[0026] The weight of each sample in the first iteration is represented as:

[0027]

[0028] Wherein, D1 represents the weight of each sample in the first iteration.

[0029] Further, the iteration module is configured to perform multiple iterations to obtain a weak classifier, wherein the weight of the weak classifier is represented as:

[0030]

[0031] Wherein, weight m represents the weight of the weak classifier; e mError function value in the last iteration.

[0032] Further, the strong classifier in the combination module is represented as:

[0033]

[0034] wherein f(x) represents the strong classifier; G m (x) represents the classification of the sample x by the weak classifier, M represents the maximum number of iterations, and sign represents the sign function.

[0035] Further, the evaluation check result in the result output module includes classified browsing of each check site image, whether there is free liquid in each check site, comprehensive evaluation of free liquid in each site, and support for clinicians to measure the volume of free liquid.

[0036] Further, the data acquisition module receives an ultrasonic echo signal related to human tissue, represented as:

[0037]

[0038] wherein f rx represents the bandwidth of the ultrasonic echo signal; f tx represents the transmission frequency of the ultrasonic probe.

[0039] Further, the texture feature of the ultrasonic image in the eFast auxiliary diagnosis module includes a feature based on a gray level co-occurrence matrix and a feature based on a gray level histogram.

[0040] Correspondingly, an automatic trauma key ultrasonic evaluation method is also provided, comprising:

[0041] S1. Receiving an ultrasonic echo signal related to human tissue, sampling the received ultrasonic echo signal through an analog-to-digital converter and forming a data stream;

[0042] S2. Receiving the data stream, demodulating the received data stream to a base frequency, and processing the signal corresponding to the data stream demodulated to the base frequency through a low-pass filter and a logarithmic compression, to obtain a processed signal;

[0043] S3. Receiving the processed signal, and using an artificial intelligence analysis algorithm of a texture feature of an ultrasonic image to divide the ultrasonic image corresponding to the processed signal into trace, small amount, medium amount and large amount according to the degree of free liquid, to obtain a diagnosis result;

[0044] S4. Receiving the processed signal, and performing scan conversion, digital gain compensation, dynamic range adjustment and image domain post-processing on the ultrasonic image corresponding to the processed signal, to obtain a processed ultrasonic image;

[0045] S5. The obtained diagnostic result and the obtained processed ultrasound image are fused to obtain an evaluation examination result corresponding to the ultrasound image.

[0046] Further, the step S3 specifically comprises:

[0047] S31. The ultrasound image is segmented into a plurality of sub-blocks, and each sub-block is independently a training sample;

[0048] S32. The weight distribution of the training sample is initialized to obtain the weight D1 of each sample in the first iteration;

[0049] S33. A plurality of iterations are performed to obtain a weak classifier; wherein the weight D of each iteration is used m The training sample set is learned, and m represents the number of iterations;

[0050] S34. When the iteration is completed, the weak classifiers are combined into a strong classifier;

[0051] S35. According to the strong classifier, the ultrasound image is divided into trace, small amount, medium amount and large amount according to the degree of free liquid, and then the severity of the injury is judged.

[0052] Compared with the prior art, the present application has the following beneficial effects:

[0053] 1. Automatically performing eFast free liquid evaluation, quickly judging chest and abdominal closed injury, providing a basis for doctors to make or adjust clinical diagnosis and treatment decisions, and improving the effectiveness of trauma ultrasound evaluation technology;

[0054] 2. Adopting an artificial intelligence training and analysis module, analyzing and classifying according to the typical features of image center bag fluid, abdominal free fluid and pelvic free fluid, and helping doctors to judge the severity of trauma.

[0055] 3. By extracting a group of texture feature parameters that can effectively distinguish between pathological and normal ultrasound images, using an Adaboost high-precision classifier, without the need for feature screening, thereby reducing the algorithm complexity, so that the system can quickly and accurately diagnose acute chest and abdominal closed injury.

[0056] 4. The eFast auxiliary diagnosis result is fused with the ultrasound imaging result to output, including classification browsing of each examination site image, whether there is free liquid in each examination site, comprehensive evaluation of free liquid in each site, and supporting the clinician to measure the volume of free liquid. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 It is a system structure diagram of an automatic trauma key ultrasound evaluation provided by Example 1.

[0058] Figure 2is a flow chart of a method for automatic trauma focus ultrasound evaluation provided in Embodiment Two. DETAILED DESCRIPTION

[0059] The present application can be implemented or applied in other different specific embodiments, and the details in the specification can be modified or changed based on different views and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0060] The purpose of the present application is to overcome the defects of the prior art, and provide a method and system for automatic trauma focus ultrasound evaluation.

[0061] Embodiment One

[0062] The present embodiment provides a system for automatic trauma focus ultrasound evaluation, as shown in Figure 1 The system comprises:

[0063] The data acquisition module 11 is configured to receive an ultrasound echo signal related to human tissue, sample the received ultrasound echo signal through an analog-to-digital converter and form a data stream, and transmit the formed data stream to the signal processing module;

[0064] The signal processing module 12 is connected with the data acquisition module 11, configured to receive the data stream, demodulate the received data stream to a base frequency, and process the signal corresponding to the data stream demodulated to the base frequency through a low-pass filter and a logarithmic compression to obtain a processed signal;

[0065] The data transmission module 13 is connected with the signal processing module 12, configured to transmit the obtained processed signal to the eFast auxiliary diagnosis module and the ultrasound imaging module;

[0066] The eFast auxiliary diagnosis module 14 is connected with the data transmission module 13, configured to receive the processed signal, and use an artificial intelligence analysis algorithm of texture features of an ultrasound image to divide the ultrasound image corresponding to the processed signal into trace, small amount, medium amount and large amount according to the degree of free liquid, and obtain a diagnosis result;

[0067] The ultrasound imaging module 15 is connected with the data transmission module 13, configured to receive the processed signal, and perform scan conversion, digital gain compensation, dynamic range adjustment and image domain post-processing on the ultrasound image corresponding to the processed signal to obtain a processed ultrasound image;

[0068] The result output module 16 is connected with the eFast auxiliary diagnosis module 14 and the ultrasonic imaging module 15 respectively, and is used for fusing the diagnosis result obtained by the eFast auxiliary diagnosis module and the processed ultrasonic image obtained by the ultrasonic imaging module to obtain an evaluation examination result corresponding to the ultrasonic image.

[0069] The system for automatic trauma key ultrasonic evaluation provided in the embodiment aims at: the trauma such as pneumothorax, hemothorax or cardiac tamponade changes the gas and liquid inside the human body, thereby producing different ultrasonic characteristics, and the automatic evaluation process utilizes the ultrasonic images with the characteristics to train an artificial intelligence recognition and classification algorithm, which can be used for evaluation and auxiliary diagnosis of emergency trauma. The process can be well combined with an existing ultrasonic system to meet the needs of clinicians to quickly and accurately evaluate acute chest and abdominal closed injuries and effectively carry out follow-up diagnosis and treatment.

[0070] In the data acquisition module 11, the ultrasonic echo signal related to the human tissue is received, the received ultrasonic echo signal is sampled by an analog-to-digital converter and a data stream is formed, and the formed data stream is transmitted to the signal processing module.

[0071] The ultrasonic echo signal of the human tissue is received by the data acquisition module, the ultrasonic echo signal is sampled by an analog-to-digital converter (such as an AFE5808 analog front-end chip), the AFE5808 analog front-end chip forms a data stream from the sampled ultrasonic echo signal and transmits the data stream to the signal processing module.

[0072] The embodiment assumes that the ultrasonic probe emits a frequency of f tx , and the ultrasonic echo signal bandwidth is f rx , and according to experience, the following can be obtained:

[0073]

[0074] According to the Nyquist theorem, the sampling rate of the AFE5808 analog front-end chip can be obtained as follows:

[0075]

[0076] wherein f sample represents the sampled ultrasonic echo signal.

[0077] In the signal processing module 12, the data stream is received, the received data stream is demodulated to a base frequency, and the signal corresponding to the data stream demodulated to the base frequency is processed by a low-pass filter and a logarithmic compression to obtain a processed signal.

[0078] The signal processing module performs beam synthesis and beam signal demodulation on the data stream, and demodulates the signal to a base frequency; and then removes the noise in the signal and retains the effective signal by low-pass filtering and logarithmic compression.

[0079] The low-pass filter of the embodiment adopts a complex Chebyshev filter, and the filter bandwidth changes synchronously with the signal bandwidth along the image depth direction. The filtered signal is further resampled to reduce the data transmission rate under the premise of meeting the Nyquist theorem.

[0080] In the data transmission module 13, the obtained processed signal is transmitted to the eFast auxiliary diagnosis module and the ultrasonic imaging module.

[0081] In the embodiment, the processed signal is transmitted into the ARM processor of the host computer through PCIe (peripheral component interconnect express) for eFast auxiliary diagnosis and ultrasonic imaging processing.

[0082] In the eFast auxiliary diagnosis module 14, the processed signal is received, and the ultrasonic image corresponding to the processed signal is divided into trace, small amount, medium amount and large amount according to the free degree of liquid by using the artificial intelligence analysis algorithm based on the texture features of the ultrasonic image, and the diagnosis result is obtained.

[0083] The eFast auxiliary diagnosis module adopts an artificial intelligence analysis algorithm based on the texture features of the ultrasonic image. The texture features of the ultrasonic image adopt features based on a gray level co-occurrence matrix and features based on a gray level histogram. The features based on the gray level co-occurrence matrix include angular second moment, contrast, variance, correlation, inverse difference moment and entropy. The features based on the gray level histogram include mean, variance, kurtosis, skewness, energy and entropy. A classic Adaboost (Adaptive Boosting) algorithm is used to combine multiple weak classifiers into a strong classifier. In each iteration process, the weight of the sample misclassified by the previous weak classifier is strengthened, and the sample with updated weight is used to train the next new weak classifier. In each round of training, the new weak classifier is trained with the overall training sample to generate new sample weight and the speaking right of the weak classifier, and the iteration is continued until the predetermined error rate is reached or the specified maximum iteration number is reached.

[0084] Specifically, it includes:

[0085] The segmentation module is used for segmenting the ultrasonic image into a plurality of sub-blocks, and each sub-block is independently a training sample;

[0086] The initialization module is used for initializing the weight distribution of the training sample to obtain the sample weight D1 of the first iteration;

[0087] In the present and strength, assuming that there are N training sample ultrasonic images, the initial weight w 1i is expressed as:

[0088]

[0089] wherein N represents the images of the training samples;

[0090] The weight of each sample in the first iteration is denoted as:

[0091]

[0092] wherein D1 represents the weight of each sample in the first iteration

[0093] The iteration module is configured to perform multiple iterations to obtain the weak classifier, wherein the weight D is used in each iteration. m The training sample set is used for learning, and m represents the number of iterations.

[0094] In the multiple iterations, the weight distribution of the training sample set is also updated for the next iteration.

[0095] The weight D is used in each iteration. m The training sample set is used for learning to obtain the weak classifier, wherein the weight of the weak classifier is weight m , which is denoted as:

[0096]

[0097] wherein weight m represents the weight of the weak classifier, and the weight represents the importance of the weak classifier in the strong classifier; e m represents the error function value in the last iteration, and the error function is minimized, that is, the sum of the weights of the misclassified samples is minimized.

[0098] The combination module is configured to combine the weak classifiers into the strong classifier when the iteration is completed.

[0099] The strong classifier is denoted as:

[0100]

[0101] wherein f(x) represents the strong classifier; G m (x) represents the classification of the sample x by the weak classifier, M represents the maximum number of iterations, and sign represents the sign function.

[0102] The classification module is configured to classify the two-dimensional signals collected according to the strong classifier f(x), and divide the ultrasound images of each part into trace, small amount, medium amount and large amount according to the degree of free liquid, and further judge the severity of the injury.

[0103] In this embodiment, the strong classifier f(x) obtained by training is used to classify the collected two-dimensional signals, and the ultrasound images of each part are divided into trace, small amount, medium amount and large amount according to the degree of free liquid, and further judge the severity of the injury.

[0104] In the ultrasound imaging module 15, the processed signal is received, and the ultrasound image corresponding to the processed signal is scan-converted, digitally gain-compensated, dynamically range-adjusted, and post-processed in the image domain to obtain a processed ultrasound image.

[0105] The ultrasound imaging module includes scan conversion, digital gain compensation, and dynamic range adjustment. After the ultrasound image is processed by scan conversion, digital gain compensation, and dynamic range adjustment, the processed ultrasound image is post-processed in the image domain, which can enhance the image edge and reduce speckle noise interference.

[0106] In the result output module 16, the diagnostic result obtained by the eFast auxiliary diagnosis module and the processed ultrasound image obtained by the ultrasound imaging module are fused to obtain an evaluation examination result corresponding to the ultrasound image.

[0107] The result output module fuses the eFast auxiliary diagnosis result and the ultrasound imaging result, delineates the free fluid region on the ultrasound image, and displays the degree of fluid separation (trace, small amount, medium amount, and large amount) on the ultrasound image. The output result includes classified browsing of images of each examination site, whether there is free fluid in each examination site, comprehensive evaluation of free fluid in each site, and support for clinicians to measure the volume of free fluid.

[0108] Compared with the prior art, the embodiment has the following beneficial effects:

[0109] 1. The eFast free fluid evaluation is automatically performed, the acute thoracic and abdominal closed injury is quickly judged, the basis for the doctor to make or adjust the clinical diagnosis and treatment decision is provided, and the effectiveness of the trauma ultrasound evaluation technology is improved.

[0110] 2. The artificial intelligence training and analysis module is adopted, the typical features of image center bag fluid, abdominal free fluid, and pelvic free fluid are analyzed and classified, and the severity of the trauma is helped to be judged by the doctor.

[0111] 3. A group of texture feature parameters capable of effectively distinguishing between lesions and normal ultrasound images are extracted, an Adaboost high-precision classifier is adopted, feature screening is not required, the algorithm complexity is reduced, and the system can quickly and accurately diagnose acute thoracic and abdominal closed injury.

[0112] 4. The eFast auxiliary diagnosis result and the ultrasound imaging result are fused and output, including classified browsing of images of each examination site, whether there is free fluid in each examination site, comprehensive evaluation of free fluid in each site, and support for clinicians to measure the volume of free fluid.

[0113] Embodiment two

[0114] The embodiment provides a method for automatically evaluating trauma key ultrasound, as shown in the following formula: Figure 2

[0115] S1. receiving an ultrasonic echo signal related to human tissue, sampling the received ultrasonic echo signal through an analog-to-digital converter and forming a data stream;

[0116] S2. receiving the data stream, demodulating the received data stream to a base frequency, and processing the signal corresponding to the data stream demodulated to the base frequency through a low-pass filter and a logarithmic compression, to obtain a processed signal;

[0117] S3. receiving the processed signal, and using an artificial intelligence analysis algorithm of texture features of an ultrasonic image to divide the ultrasonic image corresponding to the processed signal into trace, small amount, medium amount and large amount according to the degree of free liquid, to obtain a diagnosis result;

[0118] S4. receiving the processed signal, and performing scan conversion, digital gain compensation, dynamic range adjustment and post-processing in the image domain on the ultrasonic image corresponding to the processed signal, to obtain a processed ultrasonic image;

[0119] S5. fusing the obtained diagnosis result and the obtained processed ultrasonic image to obtain an evaluation examination result corresponding to the ultrasonic image.

[0120] Further, the step S3 specifically comprises:

[0121] S31. dividing the ultrasonic image into a plurality of sub-blocks, and each sub-block is independently a training sample;

[0122] S32. initializing the weight distribution of the training sample to obtain the weight D1 of each sample in the first iteration;

[0123] S33. performing multiple iterations to obtain a weak classifier; wherein the weight D m training sample set, and m represents the number of iterations;

[0124] S34. when the iteration is completed, combining the weak classifier into a strong classifier;

[0125] S35. dividing the ultrasonic image into trace, small amount, medium amount and large amount according to the degree of free liquid according to the strong classifier, and further judging the severity of the injury.

[0126] It should be noted that the method for automatically evaluating trauma key ultrasound provided in the embodiment is similar to that in Embodiment One, and will not be described here.

[0127] Compared with the prior art, the present application has the following beneficial effects:

[0128] ​1.eFAST automatic trauma assessment improves the effectiveness of existing trauma ultrasound assessment technology, which can quickly and automatically identify free fluid in the human body, judge chest and abdominal closed injury, and provide basis for doctors to make or adjust clinical diagnosis and treatment decisions;

[0129] 2. Help clinicians conduct a comprehensive ultrasound assessment of patients and draw objective conclusions to provide objective basis for follow-up diagnosis and treatment.

[0130] 3. Adopt artificial intelligence training and analysis module, analyze and classify according to the typical characteristics of image center bag liquid, abdominal free fluid and pelvic free fluid, help doctors judge the severity of trauma.

[0131] 4. By extracting a group of texture feature parameters that can effectively distinguish between pathological and normal ultrasound images, using Adaboost high-precision classifier, without the need for feature screening, thus reducing the complexity of the algorithm, so that the system can quickly and accurately diagnose acute chest and abdominal closed injury.

[0132] Note that the above is only the preferred embodiment of the present application and the technical principle applied. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and those skilled in the art can make various obvious changes, readjustments and substitutions without departing from the scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A system for automated assessment of ultrasound of a wound focus, characterized by, The application relates to an eFast auxiliary diagnosis system for free fluid in ultrasound images. The data acquisition module is configured to receive an ultrasonic echo signal related to human tissue, sample the received ultrasonic echo signal through an analog-to-digital converter, and obtain, according to the Nyquist theorem, a sampled ultrasonic echo signal, wherein the sampled ultrasonic echo signal has a bandwidth of wherein, represents the sampled ultrasonic echo signal, the ultrasonic echo signal has a bandwidth of and forms a data stream, and transmit the formed data stream to the signal processing module. The system comprises the following modules: a signal processing module connected with the data acquisition module, which is used for receiving a data stream, demodulating the received data stream to a base frequency, and obtaining a processed signal by processing a signal corresponding to the data stream demodulated to the base frequency through a low-pass filter and a logarithmic compression process, wherein the low-pass filter adopts a complex Chebyshev filter, the filter bandwidth changes synchronously along an image depth direction with the signal bandwidth, and the filtered signal is further reduced in data transmission rate through resampling under the premise of satisfying the Nyquist theorem; a data transmission module connected with the signal processing module, which is used for transmitting the obtained processed signal to an eFast auxiliary diagnosis module and an ultrasound imaging module; the eFast auxiliary diagnosis module connected with the data transmission module, which is used for receiving the processed signal, and dividing an ultrasound image corresponding to the processed signal into trace, small amount, medium amount and large amount according to the free degree of liquid by adopting an artificial intelligence analysis algorithm of texture features of the ultrasound image, so as to obtain a diagnosis result, wherein the texture features include features based on a gray level co-occurrence matrix and features based on a gray level histogram, the features based on the gray level co-occurrence matrix include an angular second moment, a contrast, a variance, a correlation, a difference between moments and an entropy, and the features based on the gray level histogram include a mean, a variance, a kurtosis, a skewness, an energy and an entropy; the ultrasound imaging module connected with the data transmission module, which is used for receiving the processed signal, and performing scan conversion, digital gain compensation, dynamic range adjustment and post-processing in an image domain on the ultrasound image corresponding to the processed signal, so as to obtain a processed ultrasound image; a result output module connected with the eFast auxiliary diagnosis module and the ultrasound imaging module respectively, which is used for fusing the diagnosis result obtained by the eFast auxiliary diagnosis module and the processed ultrasound image obtained by the ultrasound imaging module, delineating a free liquid region on the ultrasound image, displaying the free degree of liquid on the ultrasound image, and obtaining an evaluation examination result corresponding to the ultrasound image. The eFast auxiliary diagnosis module specifically comprises the following modules: An initialization module is configured to initialize the weight distribution of the training samples to obtain the weight D1 of each sample in the first iteration, and initialize the weight w 1i is represented as: ， a segmentation module, which is used for segmenting the ultrasound image into a plurality of subblocks, and each subblock is an independent training sample; wherein N represents an image of the training sample; ， the weight of each sample in the first iteration is represented as: ， wherein weight m represents the speech right of the weak classifier; e m represents the error function value in the last iteration; wherein the weight D m is used in each iteration; m represents the number of iterations; an iteration module, which is used for performing multiple iterations to obtain weak classifiers, wherein the weight of the weak classifiers is represented as: ， where f(x) represents a strong classifier; G m (x) represents the classification of the sample x by the weak classifier, M represents the maximum number of iterations, and sign represents the sign function. a combination module, which is used for combining the weak classifiers into a strong classifier by adopting an Adaboost algorithm when the iterations are completed, and the strong classifier is represented as: In the process of each iteration, the weight of the sample misclassified by the previous weak classifier is strengthened, and the sample after the weight update is used to train a new weak classifier again, in each round of training, the new weak classifier is trained by using the total training sample, the new sample weight and the weight of the weak classifier are generated, and the iterations are performed until a predetermined error rate is reached or a specified maximum number of iterations is reached; a classification module, which is used for dividing the ultrasound image into trace, small amount, medium amount and large amount according to the free degree of liquid by using the strong classifier, and further judging the severity of the damage.

2. A system for automated trauma focus ultrasound assessment according to claim 1, wherein, The evaluation check result in the result output module includes classified browsing of each check site image, whether there is free liquid in each check site, comprehensive evaluation of free liquid in each site, and supporting the clinician to measure the volume of free liquid.

3. A system for automated assessment of focal ultrasound in trauma according to claim 1, wherein, The data acquisition module receives an ultrasonic echo signal related to human tissue, which is represented as: ， where f rx represents the ultrasound echo signal bandwidth; f tx represents the ultrasound probe transmit frequency.

4. A method of automated focal ultrasound assessment of trauma, characterized by, It comprises: S1. receive an ultrasonic echo signal related to human tissue, sample the received ultrasonic echo signal through an analog-to-digital converter, and according to the Nyquist theorem, it can be obtained that: wherein, represents the sampled ultrasonic echo signal, the bandwidth of the ultrasonic echo signal is , and a data stream is formed; S2. Receiving a data stream, demodulating the received data stream to a base frequency, and processing the signal corresponding to the demodulated data stream to a base frequency through a low-pass filter and a logarithmic compression process to obtain a processed signal, wherein the low-pass filter uses a complex Chebyshev filter, and the filter bandwidth changes synchronously with the signal bandwidth along the image depth direction; the filtered signal is further reduced in data transmission rate through resampling under the premise of meeting the Nyquist theorem; S3. Receiving the processed signal, and using an artificial intelligence analysis algorithm of texture features of an ultrasonic image to divide the ultrasonic image corresponding to the processed signal into trace, small amount, medium amount and large amount according to the degree of liquid separation to obtain a diagnosis result, wherein the texture features include features based on a gray level co-occurrence matrix and features based on a gray level histogram, and the features based on the gray level co-occurrence matrix include angular second moment, contrast, variance, correlation, inverse difference moment and entropy, and the features based on the gray level histogram include mean, variance, kurtosis, skewness, energy and entropy; S4. Receiving the processed signal, and performing scan conversion, digital gain compensation, dynamic range adjustment and post-processing in the image domain on the ultrasonic image corresponding to the processed signal to obtain a processed ultrasonic image; S5. Fusing the obtained diagnosis result and the obtained processed ultrasonic image to delineate the free liquid area on the ultrasonic image and display the degree of liquid separation on the ultrasonic image to obtain an evaluation check result corresponding to the ultrasonic image; The step S3 specifically comprises: S31. Dividing the ultrasonic image into a plurality of sub-blocks, each of which is an independent training sample; S32. Initialize the weight distribution of the training sample, get the weight D1 of each sample of the first iteration, initialize the weight w 1i is represented as: ， Wherein N represents the image of the training sample; The weight of each sample in the first iteration is represented as: ， S33. Multiple iterations are performed to obtain weak classifiers, and the weight of the weak classifier is represented as: ， wherein weight m represents the speech right of the weak classifier; e m represents the error function value in the last iteration; wherein the weight D m is used in each iteration; m represents the number of iterations; S34. When the iteration is completed, the Adaboost algorithm is used to combine a plurality of weak classifiers into a strong classifier, which is represented as: ， where f(x) represents a strong classifier; G m (x) represents the classification of the sample x by the weak classifier, M represents the maximum number of iterations, and sign represents the sign function. During each iteration, the weight of the sample misclassified by the previous weak classifier is strengthened, and the sample after weight update is used to train the next new weak classifier. In each round of training, the new weak classifier is trained with the overall training, and the new sample weight and the weight of the weak classifier are generated. The iteration is continued until the predetermined error rate is reached or the specified maximum number of iterations is reached; S35. According to the strong classifier, the ultrasonic image is divided into trace, small amount, medium amount and large amount according to the degree of free liquid, and the severity of the injury is further judged.

Images