A method and device for risk assessment of a stroke patient, and a medium

Abstract

The present application relates to the technical field of health management, and particularly relates to a risk assessment method, device and equipment for a stroke patient and a medium, the method comprising: acquiring video images and audio data of the stroke patient, the video images comprising facial response images of the stroke patient based on a first guide instruction and limb response images based on a second guide instruction, and the audio data being replies of the stroke patient based on a third guide instruction; determining facial paralysis recognition results and eye response recognition results based on the facial response images; determining four-limb movement recognition results based on the limb response images; determining language articulation recognition results based on the audio data; and determining a risk assessment level of the stroke patient based on the facial paralysis recognition results, the eye response recognition results, the four-limb movement recognition results and the language articulation recognition results, so that the risk of the stroke patient can be accurately and quickly assessed without considering medical environmental conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health management, and particularly relates to a risk assessment method, device and equipment for a stroke patient and a medium. BACKGROUND

[0002] When a patient has a stroke symptom, a professional needs to assess the risk of the patient. A standard assessment form, such as the NIHHS assessment scale for a stroke patient, is usually used. During the assessment process, the professional will ask the patient to make corresponding actions and responses according to each index in the scale. Then, the professional scores each item according to the patient's responses. Finally, the risk level of the patient is determined by aggregating the scores. Different professionals may have different judgments according to their experience. In particular, in a relatively remote town hospital or health room, the risk assessment of a stroke patient often lacks sufficient professionalism and standardized procedures, which may lead to incorrect assessment of the risk level and thus miss the best treatment time.

[0003] Therefore, how to accurately and quickly assess the risk of a stroke patient is a technical problem to be solved at present. SUMMARY

[0004] In view of the above problems, the present application provides a risk assessment method, device, equipment and medium for a stroke patient, which overcomes the above problems or at least partially solves the above problems.

[0005] In a first aspect, the present application provides a risk assessment method for a stroke patient, comprising:

[0006] obtaining video images and audio data of the stroke patient, wherein the video images include facial response images made by the stroke patient based on first guide instructions and limb response images made by the stroke patient based on second guide instructions, and the audio data is a reply made by the stroke patient based on third guide instructions;

[0007] determining facial paralysis recognition results and eye response recognition results based on the facial response images;

[0008] determining four-limb movement recognition results based on the limb response images;

[0009] determining language articulation recognition results based on the audio data;

[0010] determining a risk assessment level of the stroke patient based on the facial paralysis recognition results, the eye response recognition results, the four-limb movement recognition results and the language articulation recognition results.

[0011] Preferably, before determining the facial paralysis recognition results and the eye response recognition results based on the facial response images, the method further comprises:

[0012] recognize the video image by using a Haar feature classifier, and locate a face region of the stroke patient;

[0013] determine a facial reaction image based on the face region.

[0014] Preferably, determine a facial paralysis recognition result and a gaze reaction recognition result based on the facial reaction image, including:

[0015] mark a position of each facial feature of the face region based on the facial reaction image;

[0016] determine a key feature point of each facial feature based on the position of each facial feature;

[0017] connect the key feature points of each facial feature to form a point line;

[0018] determine the facial paralysis recognition result and the gaze reaction recognition result based on a floating change of the point line.

[0019] Preferably, determine the facial paralysis recognition result and the gaze reaction recognition result based on the floating change of the point line, including:

[0020] determine a change of a target region based on the floating change of the point line;

[0021] determine the facial paralysis recognition result and the gaze reaction recognition result based on the change of the target region.

[0022] Preferably, determine the facial paralysis recognition result and the gaze reaction recognition result based on the change of the target region, including:

[0023] when recognizing a facial paralysis condition, determine the facial paralysis recognition result based on a first change of a nasolabial sulcus and a second change of a lower face;

[0024] when recognizing a gaze reaction condition, determine the gaze reaction recognition result based on a condensation of eyeballs and a reaction of the eyeballs to a stimulus.

[0025] Preferably, determine a four-limb movement recognition result based on the limb reaction image, including:

[0026] provide an image guidance frame as a reference by providing the image guidance frame;

[0027] determine a judgment result by judging whether a target limb in the limb reaction image is lifted into the image guidance frame and kept for a preset time length;

[0028] determine the four-limb movement recognition result based on the judgment result.

[0029] Preferably, based on the audio data, a language construction and pronunciation recognition result is determined, comprising:

[0030] The audio data is parsed by using a natural language processing model, and a proportion of error correction data is determined.

[0031] Based on the audio data and the second guide instruction, a reply accuracy rate is determined.

[0032] Based on the proportion of error correction data and the reply accuracy rate, a language construction and pronunciation recognition result is determined.

[0033] In a second aspect, the present application further provides a risk assessment device for a stroke patient, comprising:

[0034] An acquisition module is configured to acquire video images and audio data of a stroke patient, wherein the video images comprise facial reaction images made by the stroke patient based on a first guide instruction and limb reaction images made by the stroke patient based on a second guide instruction, and the audio data is a reply made by the stroke patient based on a third guide instruction.

[0035] A first determination module is configured to determine a facial paralysis recognition result and a eye contact reaction recognition result based on the facial reaction images.

[0036] A second determination module is configured to determine a four-limb movement recognition result based on the limb reaction images.

[0037] A third determination module is configured to determine a language construction and pronunciation recognition result based on the audio data.

[0038] A fourth determination module is configured to determine a risk assessment level of the stroke patient based on the facial paralysis recognition result, the eye contact reaction recognition result, the four-limb movement recognition result and the language construction and pronunciation recognition result.

[0039] In a third aspect, the present application further provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to realize the method in the first aspect.

[0040] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program executable by a processor to realize the method in the first aspect.

[0041] The one or more technical solutions in the embodiments of the present application have at least the following technical effects or advantages:

[0042] The present invention provides a risk assessment method for stroke patients, comprising: acquiring video images and audio data of the stroke patient, the video image comprising a facial reaction image of the stroke patient based on a first guidance instruction and a limb reaction image based on a second guidance instruction, and the audio data being a reply made by the stroke patient based on a third guidance instruction; determining facial paralysis recognition results and eye reaction recognition results based on the facial reaction image; determining limb movement recognition results based on the limb reaction image; determining language articulation recognition results based on the audio data; determining language articulation recognition results based on the facial paralysis recognition result and the eye reaction recognition result; determining language articulation recognition results based on the audio data; determining the risk assessment level of the stroke patient based on the facial paralysis recognition result, the eye reaction recognition result, the limb movement recognition result, and the language articulation recognition result, thereby accurately and quickly assessing the risk of the stroke patient without considering medical environment conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. Throughout the drawings, the same reference figures denote the same components. In the drawings:

[0044] Figure 1 A schematic diagram of the steps of a risk assessment method for stroke patients according to an embodiment of the present invention is shown;

[0045] Figure 2 A schematic diagram showing the distribution of key feature points of each facial feature in an embodiment of the present invention;

[0046] Figure 3 A schematic diagram showing the position lines of key feature points of each facial feature in an embodiment of the present invention is shown;

[0047] Figure 4 A schematic diagram showing an image guidance frame diagram corresponding to a stroke patient in a sitting position according to an embodiment of the present invention is shown;

[0048] Figure 5 A schematic diagram showing an image guidance frame diagram corresponding to a stroke patient in a supine position according to an embodiment of the present invention is shown;

[0049] Figure 6 A schematic structural diagram of a risk assessment device for stroke patients according to an embodiment of the present invention is shown;

[0050] Figure 7 A schematic structural diagram of a computer device for implementing a risk assessment method for stroke patients according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0051] Exemplary embodiments of the present application will be described in greater detail below with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it is understood that the present application can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the application to those skilled in the art.

[0052] Embodiment one:

[0053] The embodiments of the present application provide a risk assessment method for a stroke patient, as shown in the following formula (I): Figure 1 The formula (I) comprises the following steps:

[0054] S101, acquiring video images and audio data of a stroke patient, the video images comprising facial reaction images of the stroke patient based on first guide instructions and limb reaction images based on second guide instructions, and the audio data being replies of the stroke patient based on third guide instructions;

[0055] S102, determining facial paralysis recognition results and eye contact recognition results based on the facial reaction images;

[0056] S103, determining four-limb movement recognition results based on the limb reaction images;

[0057] S104, determining language articulation recognition results based on the audio data;

[0058] S105, determining a risk assessment grade of the stroke patient based on the facial paralysis recognition results, the eye contact recognition results, the four-limb movement recognition results, and the language articulation recognition results.

[0059] Stroke, commonly known as apoplexy, includes ischemic stroke (also known as cerebral infarction) and hemorrhagic stroke (including intracerebral hemorrhage, intraventricular hemorrhage, and subarachnoid hemorrhage), which is a disease of cerebral cell and tissue necrosis.

[0060] Stroke has the characteristics of high incidence, high disability rate, high recurrence rate, and high mortality rate, and therefore, it needs to be paid enough attention.

[0061] Common symptoms of stroke include sudden numbness or weakness, especially abnormality of one side of the body, language or understanding problems, vision problems, headache, dizziness, etc. These are subjective feelings of the patient, and in order to be able to directly observe the patient's performance from the appearance, professional assessment of the medical and nursing end is needed, and therefore, video images and audio data of the stroke patient can be collected.

[0062] Therefore, S101 is performed to acquire video images and audio data of the stroke patient, wherein the video images include facial response images of the stroke patient based on first guiding instructions and limb response images of the stroke patient based on second guiding instructions, and the audio data is a reply of the stroke patient based on third guiding instructions.

[0063] During the acquisition of the video images and the audio data of the stroke patient, the stroke patient needs to be guided by voice to perform some actions and replies.

[0064] The first guiding instructions include guiding the stroke patient to smile.

[0065] The second guiding instructions include guiding the limbs of the stroke patient to reach appropriate positions and make corresponding actions, such as guiding the stroke patient to sit and hold the upper limbs at 90°, guiding the stroke patient to lie on the back and lift the legs at 45°, and the like.

[0066] The third guiding instructions include guiding the stroke patient to answer basic questions and the like.

[0067] Next, the acquired video images and audio data need to be analyzed. S102 is performed to determine facial paralysis recognition results and eye response recognition results based on the facial response images.

[0068] Here, the facial paralysis recognition and the eye response recognition are needed.

[0069] For the facial paralysis recognition, the following manner is specifically adopted for recognition:

[0070] Before the facial paralysis recognition, the following is further included: using a Haar feature classifier to recognize the video images and locate a face region of the stroke patient; and determining the facial response images based on the face region.

[0071] The Haar feature classifier can clearly distinguish between a face and a non-face, so as to determine the face region in the video images and obtain the facial response images in the face region.

[0072] In a specific implementation, each frame of the video images is divided into multiple regions, and each region is input into the Haar feature classifier to determine the approximate face region. When divided into multiple regions, the whole body is divided.

[0073] After the approximate face region is obtained, the non-face region is removed and the face region is retained.

[0074] Then, the facial response images of the face region are captured. Since the facial response images include facial expression conditions and eye response conditions, the facial paralysis recognition results and the eye response recognition results can be respectively determined through the facial response images.

[0075] In the process of determining these two recognition results, the following methods are used for identification:

[0076] Based on the facial response image, mark the position of each facial feature in the facial area;

[0077] Based on the position of each facial feature, determine the key feature points of each facial feature;

[0078] Connect the key feature points of each facial feature to form a point line;

[0079] Based on the floating changes of the point line, the facial paralysis recognition results and the eye reaction recognition results are determined.

[0080] like Figure 2 As shown in the figure, it shows the distribution of key feature points of each facial feature. Figure 2 In addition to the key feature points of each facial feature, the key feature points are also marked in the area between the facial features, so that the changes in the key feature point lines of the area between the facial features can be determined based on the point lines of the key feature points of each facial feature, so as to more clearly identify facial reactions. Figure 3 Shown is a schematic diagram of the point line situation of the key feature points of each facial feature.

[0081] Next, the floating changes of the point line are determined by tracking each frame of the facial reaction image.

[0082] Specifically, based on the floating changes of the point line, the change of the target area is determined; then, based on the change of the target area, the facial paralysis recognition result and the eye reaction recognition result are determined.

[0083] The target area here is related to the reaction to be recognized. For example, to recognize whether someone is crying, the focus is on the eye area, and to recognize whether someone is smiling, the focus is on the mouth area.

[0084] The following describes facial paralysis recognition and eye reaction recognition separately:

[0085] When identifying facial paralysis, a facial paralysis identification result is determined based on a first change in the nasolabial groove and a second change in the lower face.

[0086] When the stroke patient is instructed to smile during the first guidance instruction, the area from the sides of the nose to the corners of the mouth (the nasolabial folds) is determined based on the changes in the corresponding point lines of the nose and mouth. This area is also known as the nasolabial folds, as well as the area below the mouth (the lower face). The nasolabial folds and the lower face are then identified as target areas.

[0087] For the nasolabial sulcus area, first, the first change of the nasolabial sulcus is collected, and then based on the first change, it is judged whether the nasolabial sulcus area is flattened and whether the left and right sides of the nasolabial sulcus are asymmetric.

[0088] For the lower face area, first, the second change of the lower face is collected, and then based on the second change, it is judged whether the lower face area cannot make a smiling action and whether the left or right side of the lower face area has no action.

[0089] Finally, according to the judgment result, the facial paralysis recognition result is determined. For the judgment result of each item, the score of each item is obtained, and the scores of each item are added to obtain the facial paralysis recognition result, that is, the score of facial paralysis recognition.

[0090] Among them, the nasolabial sulcus is located on both sides of the alae nasi to both sides of the corners of the mouth, and is the result of the interaction of the dynamic soft tissue and the non-dynamic soft tissue of the cheek, which will be highlighted when smiling.

[0091] When the nasolabial sulcus area is flattened, it is determined that the patient has facial paralysis, and the asymmetry of the left and right sides of the nasolabial sulcus also indicates that one side of the cheek has facial paralysis.

[0092] The face is usually divided into three regions: upper, middle and lower. The region above the eyes is called the upper face, the region from the eyes to the corners of the mouth is called the middle face, and the region below is called the lower face.

[0093] When the patient smiles, if the left or right side of the lower face area has no obvious change, it is determined that the lower face area cannot make a smiling action, that is, it is determined that the face has paralysis.

[0094] If any of the above is judged to be yes, the higher the score result.

[0095] The eye response recognition is described as follows:

[0096] In identifying the eye response, the eye ball condensation and the eye ball response to the stimulus are based on.

[0097] When the first guide instruction guides the eye ball movement of the stroke patient, the eye ball area is taken as the target area according to the floating change of the eye ball point line. For the eye ball area, first, the condensation of the eye ball is collected, that is, the tracking ability of the eye ball point line corresponding to the change of the tracking target is collected. More specifically, according to the change rate of the point line of the key feature point of the eye ball, the tracking ability of the eye ball is determined.

[0098] At the same time, the reaction of the eye ball to the stimulus such as light is also collected, so as to judge whether the eye ball has paralysis and no response to the stimulus. More specifically, the reaction time to the stimulus is determined, the longer the reaction time, the higher the paralysis degree, and the shorter the reaction time, the lower the paralysis degree.

[0099] If the tracking ability of the eyeball is poor and the degree of paralysis of the stimulus is high, the eye contact reaction result is determined, and the score of the eye contact reaction is obtained.

[0100] Next, S103 is executed, and the limb movement recognition result is determined based on the limb reaction image.

[0101] Specifically, by providing an image guidance frame, the image guidance frame is used as a reference;

[0102] Determine whether the target limb in the limb reaction image is lifted into the image guidance frame and maintained for a predetermined time period to obtain a judgment result;

[0103] Based on the judgment result, the four-limb movement recognition result is determined.

[0104] When the first guidance instruction is to guide the patient to sit, the patient's upper limbs are lifted 90°, and when the patient is guided to lie on his back, the patient's legs are lifted to 45°. According to the collected limb reaction image combined with the image guidance frame, it is determined whether the target limb, i.e. the patient's upper and lower limbs, is in the image guidance frame, and the time to complete the target action is evaluated. As shown in Figure 4 、 Figure 5 .

[0105] For the reference of the image guidance frame, the proportion of the target limb of the patient located in the image guidance frame can be determined.

[0106] According to the completion of the stroke patient, on the one hand, the proportion of the target limb and the image guidance frame needs to be determined, and on the other hand, the time to complete the target action needs to be determined, and the respective weight ratios are added to obtain the evaluation result of the four-limb movement recognition. The specific evaluation result can be divided into upper limb evaluation result and lower limb evaluation result, and the final evaluation result of the four-limb movement recognition will add the upper limb evaluation result and the lower limb evaluation result.

[0107] Obviously, the lower the proportion of the target limb and the image guidance frame, and the shorter the time to complete the target action, the lower the evaluation result of the four-limb movement recognition, and vice versa.

[0108] Next, the audio data also needs to be recognized and evaluated. S104 is executed, and the language construction sound recognition result is determined based on the audio data.

[0109] Specifically, a natural language processing model is used to analyze the audio data to determine the proportion of error correction data;

[0110] Based on the audio data and the second guidance instruction, the reply accuracy rate is determined;

[0111] Based on the proportion of error correction data and the reply accuracy rate, the language construction sound recognition result is determined.

[0112] Before determining the language pronunciation recognition result, a filter is used to remove the environment or other noise mixed in the audio data. Then, the audio data after filtering out the noise is converted into text data through text conversion. Next, the error correction data proportion determination is performed.

[0113] The natural language processing module (HanLP) can correct the text data of the audio data, record the number of corrections, calculate the proportion of the number of corrections, and use the proportion of the number of corrections as a weight basis for measuring language clarity.

[0114] Then, based on the audio data and the second guide instruction, the reply accuracy rate is determined. The audio data is a reply based on the second guide instruction, for example, the second guide instruction is, how old are you, what is your address, then, the reply is recognized, the reply to the age is compared with the true situation, the reply to the address is compared with the actual situation, the comparison result is obtained, of course, there are other question and answer processes, which are not limited here. Based on the comparison result, the reply accuracy rate is determined. The higher the accuracy rate, the lower the score, the lower the accuracy rate, the higher the score.

[0115] The evaluation results of the error correction data proportion and the evaluation results of the reply accuracy rate are combined to obtain the language pronunciation recognition result.

[0116] Specifically, the evaluation results of the error correction data proportion and the evaluation results of the reply accuracy rate are added to obtain the language pronunciation recognition result. The language pronunciation recognition result evaluates the audio data from the accuracy and clarity, and can more accurately reflect the recognition result of the audio data.

[0117] After face recognition, body recognition, and audio recognition, S105 is executed to determine the risk assessment level of the stroke patient based on the facial paralysis recognition result, the eye contact reaction recognition result, the four-limb movement recognition result, and the language pronunciation recognition result.

[0118] The scores of the facial paralysis recognition result, the eye contact reaction recognition result, the four-limb movement recognition result, and the language pronunciation recognition result can be added in sequence to obtain a total score value, and then the total score value is corresponded to the risk assessment level to determine the corresponding risk assessment level. For example, the risk assessment level has low risk, medium risk, and high risk. The higher the total score value, the higher the risk assessment level of the stroke patient, and the lower the total score value, the lower the risk assessment level of the stroke patient.

[0119] With such an evaluation method, most non-professionals can assist in the evaluation, and thus in medical conditions relatively poor township health centers, the criticality of patients can also be quickly identified, so as to ensure that patients can get timely treatment conditions.

[0120] The one or more technical solutions in the embodiments of the present application have at least the following technical effects or advantages:

[0121] The present application provides a kind of risk assessment method of stroke patient, comprising: obtaining the video image and audio data of stroke patient, video image includes the facial response image of stroke patient based on first guide instruction and the limb response image based on second guide instruction, audio data is the reply of stroke patient based on third guide instruction;Based on facial response image, determine facial paralysis recognition result and eye contact response recognition result;Based on limb response image, determine four-limb movement recognition result;Based on audio data, determine language phonation recognition result;Based on facial paralysis recognition result, eye contact response recognition result, four-limb movement recognition result, language phonation recognition result, determine the risk assessment grade of stroke patient, and then without considering medical environment condition, the risk of stroke patient can be accurately and quickly evaluated, improve diagnosis efficiency.

[0122] Embodiment two:

[0123] Based on the same inventive concept, the present application also provides a kind of risk assessment device of stroke patient, as shown in Figure Figure 6 It includes:

[0124] The acquisition module 601 obtains the video image and audio data of stroke patient, and the video image includes the facial response image of stroke patient based on first guide instruction and the limb response image based on second guide instruction, and the audio data is the reply of stroke patient based on third guide instruction;

[0125] The first determination module 602 determines facial paralysis recognition result and eye contact response recognition result based on the facial response image;

[0126] The second determination module 603 determines four-limb movement recognition result based on the limb response image;

[0127] The third determination module 604 determines language phonation recognition result based on the audio data;

[0128] The fourth determination module 605 determines the risk assessment grade of stroke patient based on the facial paralysis recognition result, eye contact response recognition result, four-limb movement recognition result, language phonation recognition result.

[0129] In an optional implementation, further comprising: a fifth determination module for:

[0130] recognize the video image by using a Haar feature classifier, and locate a face region of the stroke patient;

[0131] determine a facial response image based on the face region.

[0132] In an optional implementation, the first determining module 602 is configured to:

[0133] mark a position of each facial feature of the face region based on the facial response image;

[0134] determine a key feature point of each facial feature based on the position of each facial feature;

[0135] concatenate the key feature points of each facial feature to form a point line;

[0136] determine the facial paralysis recognition result and the eye response recognition result based on a floating change of the point line.

[0137] In an optional implementation, the first determining module 602 is configured to:

[0138] determine a change of a target region based on a floating change of the point line;

[0139] determine the facial paralysis recognition result and the eye response recognition result based on the change of the target region.

[0140] In an optional implementation, the first determining module 602 is configured to:

[0141] when recognizing the facial paralysis, determine the facial paralysis recognition result based on a first change of a nasolabial sulcus and a second change of a lower face;

[0142] when recognizing the eye response, determine the eye response recognition result based on a condensation of eyeballs and a reaction of the eyeballs to a stimulus.

[0143] In an optional implementation, the second determining module 603 is configured to:

[0144] provide an image guidance frame, and use the image guidance frame as a reference;

[0145] determine whether a target limb in the limb response image is lifted into the image guidance frame and kept for a preset time length, to obtain a determination result;

[0146] determine the four-limb movement recognition result based on the determination result.

[0147] In an optional implementation, the third determining module 604 is configured to:

[0148] The natural language processing model is used to analyze the audio data to determine a proportion of error correction data;

[0149] Based on the audio data and the second guidance instruction, a reply accuracy rate is determined;

[0150] Based on the proportion of error correction data and the reply accuracy rate, a language construction and pronunciation recognition result is determined.

[0151] Embodiment three:

[0152] Based on the same inventive concept, the present application provides a computer device, as shown in the accompanying drawings, comprising a memory 704, a processor 702, and a computer program stored on the memory 704 and executable on the processor 702, wherein the processor 702 executes the program to implement the steps of the risk assessment method for stroke patients described above. Figure 7

[0153] In the above embodiment, the bus architecture (represented by bus 700), the bus 700 can include any number of interconnected buses and bridges, the bus 700 links various circuits including one or more processors represented by processor 702 and memory represented by memory 704. The bus 700 can also link various other circuits such as peripheral devices, voltage stabilizers and power management circuits, etc., which are well known in the art, therefore, they will not be further described herein. The bus interface 706 provides an interface between the bus 700 and the receiver 701 and the transmitter 703. The receiver 701 and the transmitter 703 can be the same element, i.e. a transceiver, which provides a unit for communicating with various other devices on a transmission medium. The processor 702 is responsible for managing the bus 700 and general processing, while the memory 704 can be used to store data used by the processor 702 in performing operations. Figure 7

[0154] Embodiment four:

[0155] Based on the same inventive concept, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the steps of the risk assessment method for stroke patients described above.

[0156] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems can be used with these teachings, with or without accompanying software. Those skilled in the art will recognize that structures embodied by these descriptions might be subjected to numerous modifications, and yet be enabled to provide the functionality described herein. In addition, the present application is not intended to be limited to a particular programming language. It will be appreciated that there are many programming languages that can be used to implement the teachings described herein, and any such programming language can be used in connection with the various aspects of the application. The descriptions above are intended to provide specific examples of the application and should not be construed as limiting the scope of the application as set forth in the appended claims. ​​

[0157] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been described in detail in order not to obscure the understanding of this description.

[0158] Similarly, it is to be understood that the embodiments of the present application can be readily combined with one another, and the various features from one embodiment can be incorporated into another embodiment. In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been described in detail in order not to obscure the understanding of this description.

[0159] Those skilled in the art will appreciate that modules in the apparatuses in the embodiments can be adapted and placed in one or more apparatuses other than the embodiments. Modules or units or components in the embodiments can be combined into one module or unit or component and further can be divided into more sub-modules or sub-units or sub-components. Any combination of all or some of the disclosed features and any method or device of any of the disclosed features can be used in any combination with each other. Unless explicitly stated otherwise, each feature disclosed in the specification (including the claims, abstract and drawings) can be replaced by alternative features that serve the same, equivalent or similar purpose.

[0160] Further, those skilled in the art will appreciate that, although some embodiments herein include certain features of other embodiments but not others, combinations of the features of the different embodiments are to be expected and are within the scope of the application and form different embodiments. For example, in the DETAILED DESCRIPTION, any of the claimed embodiments can be used in any combination.

[0161] The various component embodiments of the present application can be implemented in hardware, or as software modules running in one or more processors, or in combinations thereof. As will be appreciated by one skilled in the art, microprocessors or digital signal processors (DSPs) can be used to implement some or all of the functionality of some or all of the components of the stroke risk assessment device for stroke patients, computer device according to the embodiments of the present application in practice. The present application can also be implemented as a program for executing, in whole or in part, the methods described herein on a device or apparatus (e.g., computer program and computer program product). Such program implementing the present application can be stored on a computer readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or in any other form.

[0162] It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that one skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word 'comprising' does not exclude the presence of elements or steps other than those listed in a claim. The word 'a' or 'an' preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of both hardware and software, and any combination thereof. In a unit claim, several devices can be listed with a conjunction like 'or', but it is to be understood that each of these devices can be implemented by its own hardware item. The use of the word 'at least' followed by a list of one or more members does not preclude the existence of additional such members. The use of the words 'first' and'second' does not preclude the existence of three or more such members.

Claims

1. A method for risk assessment of stroke patients, characterized in that: include: Acquiring video images and audio data of a stroke patient, wherein the video images include facial reaction images of the stroke patient based on the first guidance instruction and limb reaction images based on the second guidance instruction, and the audio data includes responses of the stroke patient based on the third guidance instruction; Determining a facial paralysis recognition result and an eye reaction recognition result based on the facial reaction image includes: marking a position of each facial feature in the facial region based on the facial response image; Based on the position of each facial feature, determine the key feature points of each facial feature; Connect the key feature points of each facial feature to form a point line; Based on the floating changes of the point line, the facial paralysis recognition results and the eye reaction recognition results are determined; The method of determining the facial paralysis recognition result and the eye reaction recognition result based on the floating change of the point line includes: Determine the changes in the target area based on the floating changes of the point line; Determining facial paralysis recognition results and eye expression recognition results based on changes in the target area; Determining facial paralysis recognition results and eye reaction recognition results based on changes in the target area includes: When identifying facial paralysis, determining a facial paralysis identification result based on a first change in the nasolabial groove and a second change in the lower face; When identifying eye reactions, the eye reaction recognition result is determined based on the eye's concentration and the eye's reaction to the stimulus; Determining limb movement recognition results based on the limb reaction image, including: By providing an image guide frame, the image guide frame is used as a control; Determine whether the target limb in the limb reaction image is lifted into the image guidance frame and maintained for a preset time period, and obtain a determination result; Determining limb movement recognition results based on the judgment result; Determining a speech articulation recognition result based on the audio data includes: Analyze the audio data using a natural language processing model to determine the proportion of error-corrected data; determining a response accuracy rate based on the audio data and the second guidance instruction; Determining a speech articulation recognition result based on the error correction data ratio and the response accuracy rate; Based on the facial paralysis recognition results, eye reaction recognition results, limb movement recognition results, and language articulation recognition results, a risk assessment level of the stroke patient is determined, including: The facial paralysis recognition results, eye reaction recognition results, limb movement recognition results, and language articulation recognition results are added together in sequence to obtain an overall score. Match the overall score value range with the risk assessment level to determine the corresponding risk assessment level.

2. The method according to claim 1, wherein Before determining the facial paralysis recognition result and the eye reaction recognition result based on the facial reaction image, the method further includes: Using a Haar feature classifier to identify the video image and locate the facial area of ​​the stroke patient; Based on the face region, a facial reaction image is determined.

3. A risk assessment device for stroke patients, characterized in that: include: an acquisition module, configured to acquire video images and audio data of a stroke patient, wherein the video images include facial reaction images of the stroke patient based on the first guidance instruction and limb reaction images based on the second guidance instruction, and the audio data are responses of the stroke patient based on the third guidance instruction; A first determination module is configured to determine a facial paralysis recognition result and an eye reaction recognition result based on the facial reaction image. The first determination module is configured to: marking a position of each facial feature in the facial region based on the facial response image; Based on the position of each facial feature, determine the key feature points of each facial feature; Connect the key feature points of each facial feature to form a point line; Based on the floating changes of the point line, the facial paralysis recognition results and the eye reaction recognition results are determined; The first determining module is configured to: Determine the changes in the target area based on the floating changes of the point line; Determining facial paralysis recognition results and eye expression recognition results based on changes in the target area; The first determining module is configured to: When identifying facial paralysis, determining a facial paralysis identification result based on a first change in the nasolabial groove and a second change in the lower face; When identifying eye reactions, the eye reaction recognition result is determined based on the eye's concentration and the eye's reaction to the stimulus; The second determination module is configured to determine a limb movement recognition result based on the limb reaction image, and the second determination module is configured to: By providing an image guide frame, the image guide frame is used as a control; Determine whether the target limb in the limb reaction image is lifted into the image guidance frame and maintained for a preset time period, and obtain a determination result; Determining limb movement recognition results based on the judgment result; A third determination module is configured to determine a speech articulation recognition result based on the audio data, wherein the third determination module is configured to: Analyze the audio data using a natural language processing model to determine the proportion of error-corrected data; determining a response accuracy rate based on the audio data and the second guidance instruction; Determining a speech articulation recognition result based on the error correction data ratio and the response accuracy rate; The fourth determination module is used to determine the risk assessment level of the stroke patient based on the facial paralysis recognition results, eye reaction recognition results, limb movement recognition results and language articulation recognition results.

4. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to claim 1 or 2 is implemented.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to claim 1 or 2 is implemented.

Images