CN115376519A - Method and equipment for generating electronic medical record and computer readable storage medium - Google Patents

Abstract

A method, apparatus, and computer-readable storage medium for generating an electronic medical record are disclosed. The method comprises the following steps: acquiring audio data related to inquiry from a left channel and a right channel, wherein the left channel and the right channel are respectively provided with respective role labels; performing voice recognition on the audio data related to the inquiry so as to at least obtain a dialog text corresponding to the character tag; respectively extracting key information related to the inquiry on the basis of the dialog text; and generating the electronic medical record according to the role labels and the key information in the corresponding dialogue text by using the neural network model. By using the scheme of the application, the audio data can respectively correspond to the medical staff and the patient, so that errors of the electronic medical record report can be avoided.

Description

A method, device and computer-readable storage medium for generating electronic medical records

technical field

This application generally relates to the technical field of electronic medical records. More specifically, the present application relates to a method, device and computer-readable storage medium for generating electronic medical records.

Background technique

Medical records are records of medical activities such as inspection, diagnosis, and treatment of the occurrence, development, and outcome of a patient's disease by medical staff during the consultation process. Format and requirements for writing patient medical records.

In the traditional consultation process, medical personnel input consultation information through keyboard input to generate a medical record report, but this will result in a long time-consuming consultation process. At present, speech recognition technology has been applied to the process of medical consultation. By identifying the audio data of the medical consultation, the identified information is filled in the corresponding position, thereby generating a medical record report. However, the audio data of the current consultation often cannot correspond to the medical staff and patients, resulting in errors when filling the identified information into the corresponding positions. For example, the question information of the medical staff is filled in the description information of the patient, and the description information of the patient is filled in the question information of the medical staff, so that the generated medical record report is wrong. In addition, the existing electronic medical record generation often adopts a fixed electronic medical record template, and the electronic medical record is generated by obtaining the corresponding relationship between the key information and the template, which makes the form of the generated electronic medical record relatively fixed and single.

Contents of the invention

In order to at least partly solve the technical problems mentioned in the background art, the solution of the present application provides a solution for generating electronic medical records. Using the solution of the present application, the audio data can be respectively corresponding to the medical personnel and patients, and the electronic medical record can be generated through the neural network model, so as to avoid errors in the electronic medical record report and generate a single form of electronic medical record report. To this end, the present application provides solutions in the following aspects.

In the first aspect, the present application provides a method for generating an electronic medical record, including: obtaining audio data related to consultation from the left and right channels, wherein the left and right channels are respectively provided with their own character label; perform speech recognition on the audio data related to the consultation, so as to at least obtain the dialogue text corresponding to the role label; extract the key information related to the consultation based on the dialogue text; and use the neural network model according to the role The key information in the label and the corresponding dialogue text generates an electronic medical record.

In one embodiment, the method further includes: in response to the stereo input of the character, setting the stereo input of the character to a stereo format; and performing channel splitting on the stereo format to determine the left channel and the right channel.

In another embodiment, wherein the role tag includes a doctor tag and a patient tag, the dialogue text includes a medical question text corresponding to the doctor tag and a medical question answer text corresponding to the patient tag.

In yet another embodiment, extracting the key information related to the medical inquiry based on the dialogue text includes: respectively marking the text of the medical question and the text of the answer; and according to the text of the marked medical question and the text of the question answer to extract their key information.

In yet another embodiment, wherein the key information includes at least the characteristic words and/or sentences about the patient's chief complaint question asked by the doctor and the characteristic words and/or sentences of the disease question and the characteristic words and/or sentences about the chief complaint described by the patient and Symptoms describe characteristic words and/or sentences.

In yet another embodiment, wherein using the neural network model to generate the electronic medical record according to the key information in the role label and the corresponding dialogue text includes: inputting the role label and the key information in the corresponding dialogue text into the neural network model to obtain A picture associated with the medical record; and generating an electronic medical record based on the picture.

In yet another embodiment, the method further includes: performing speech recognition on the audio data related to the consultation to obtain voiceprint information of the left channel and the right channel; and The pattern information determines whether the left channel and the right channel correspond to their respective role labels.

In yet another embodiment, the method further includes: providing the character with question prompt information and answer prompt information according to the key information in the character label and the corresponding dialogue text.

In yet another embodiment, providing the character with question prompt information and answer prompt information according to the key information in the corresponding dialogue text includes: according to the key information in the corresponding dialogue text, using the knowledge map related to the consultation to calculate its correspondence value; and in response to the corresponding value hitting the path of the knowledge graph, providing question prompt information and answer prompt information for the corresponding character.

In the second aspect, the present application provides a device for generating electronic medical records, including: a processor; and a memory, which stores program instructions for generating electronic medical records, when the program instructions are executed by the processor, so that The device implements several of the aforementioned embodiments.

In the third aspect, the present application provides a computer-readable storage medium on which computer-readable instructions for generating electronic medical records are stored. When the computer-readable instructions are executed by one or more processors, the aforementioned multiple Example.

Through the solution of this application, the audio data related to the consultation is obtained through the left and right channels with corresponding role labels, and by identifying and extracting key information from the respective audio data of the left and right channels, The neural network model is used to generate electronic medical records based on the aforementioned key information. Based on this, the corresponding roles (including doctors and patients) can be distinguished, and each audio data can be associated with role labels to generate correct electronic medical record reports. Further, in the embodiment of the present application, the key information is processed into pictures by using the neural network model, so that the medical staff can arrange the pictures according to the needs, and customize the form of the electronic medical record, thereby avoiding the generation of a fixed and single electronic medical record. Furthermore, the embodiment of the present application also recognizes the voiceprint information of the left and right channels to correct in time when the audio data of the left and right channels are inconsistent with the role label, thereby ensuring that the electronic medical records Reporting Accuracy. In addition, the embodiment of the present application can also prompt the doctor's question and the patient's answer, so as to improve the efficiency of consultation.

Description of drawings

The above and other objects, features and advantages of the exemplary embodiments of the present application will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present application are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

Fig. 1 is an exemplary flowchart showing a method for generating an electronic medical record according to an embodiment of the present application;

FIG. 2 is an exemplary schematic diagram illustrating obtaining dialogue text corresponding to a role label according to an embodiment of the present application;

Fig. 3 is an exemplary schematic diagram illustrating generating an electronic medical record report according to an embodiment of the present application;

FIG. 4 is an exemplary schematic diagram showing an electronic medical record report generated according to an embodiment of the present application;

Fig. 5 is an exemplary schematic diagram showing an electronic medical record report generating errors according to an embodiment of the present application;

Fig. 6 is an exemplary schematic diagram illustrating providing question prompt information and answer prompt information according to an embodiment of the present application; and

Fig. 7 is a block diagram illustrating an exemplary structure of a device for generating an electronic medical record according to an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings. It should be understood that the embodiments described in this specification are only some of the embodiments provided by the present application to facilitate a clear understanding of the solution and comply with legal requirements, but not all the embodiments of the present application can be implemented. Based on the embodiments disclosed in this specification, all other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of the present application.

Fig. 1 is an exemplary flowchart showing a method 100 for generating an electronic medical record according to an embodiment of the present application. As shown in FIG. 1 , at step S102 , audio data related to the consultation is acquired from the left and right channels. In one embodiment, the aforementioned left channel and right channel can be determined through the following operations, that is, in response to the stereo input of the character, the stereo input of the character is set to a stereo format, and then the stereo format is divided into channels to determine left and right channels. Among them, the aforementioned roles include doctors and patients. In the implementation scenario, two voice input devices (such as microphones, earphones, etc.) can be set, and the doctor and the patient can ask and answer questions through their respective voice input devices to input stereo. In this scenario, after the doctor and the patient input stereo, they can set their respective stereo input to stereo format through, for example, the API interface provided by MDN. Then, it can be divided by using, for example, a node (node), so as to be divided into a left channel and a right channel. Wherein, the aforementioned left and right channels are respectively provided with respective role tags, and the role tags may include doctor tags and patient tags. That is, the left and right channels correspond to doctors and patients, respectively. Thereby, audio data of doctors and patients can be distinguished.

Next, at step S104, speech recognition is performed on the audio data related to the consultation, so as to at least obtain dialogue text corresponding to the role tag. In an implementation scenario, speech recognition may be performed on the audio data by, for example, speech recognition software or a speech recognition program, so as to obtain dialogue text corresponding to the character tag. That is, the dialogue text obtained through recognition corresponds to the character label respectively. The present application does not limit the aforementioned speech recognition software or speech recognition program, as long as the dialogue text information can be recognized, all belong to the protection scope of the present application. Wherein, the aforementioned dialogue text may include question texts corresponding to doctor tags and question answer texts corresponding to patient tags, the question texts asked by doctors correspond to doctor tags, and the question answer texts answered by patients correspond to patient tags. In the implementation scenario, after speech recognition, the doctor's label will be in front of the question text asked by the doctor, and the patient label will be in front of the answer text of the patient's answer. For example, in an exemplary scenario, the doctor and the patient output audio data via the left channel and the right channel respectively, and after voice recognition, obtain, for example, "Doctor: What symptoms do you have in the eyes? Patient: There is a foreign body sensation in the eyes".

Based on the dialogue text obtained above, at step S106, the key information related to the consultation is extracted based on the dialogue text. In one embodiment, the medical question text and the medical question answer text can be marked separately, and then the respective key information can be extracted according to the marked medical question text and the medical question answer text. Specifically, the medical question text and the medical question answer text can be marked respectively by using, for example, an entity labeling model. In the application scenario, the entity labeling model can be trained first through a large number of disease question texts and answer texts and their corresponding entity labels, so that the entity labeling model can extract key information related to the consultation. In some embodiments, the aforementioned key information may include the characteristic words and/or sentences of the patient's chief complaint question and the characteristic words and/or sentences of the disease question asked by the doctor, the characteristic words and/or sentences of the chief complaint description and the disease description characteristics answered by the patient words and/or sentences. Among them, the chief complaint of the aforementioned patient may include but not limited to the age, address, disease history, etc. , redness, or congestion. In addition, it may also include the onset time of the disease, whether to take medication, etc. Therefore, when training the entity labeling model, it is possible to label, for example, age, eye disease, redness, foreign body sensation, etc., to output key information. For example, after the physical labeling model can output age, 22, eye symptoms, foreign body sensation in the eyes, and tingling, itching, etc.

Further, at step S108, the neural network model is used to generate an electronic medical record according to the key information in the role label and the corresponding dialogue text. In one embodiment, the key information in the character label and the corresponding dialogue text can be input into the neural network model to obtain pictures related to the medical records, and then the electronic medical records are generated based on the pictures. As known from the description of background technology, the generation of existing electronic medical records often adopts fixed electronic medical record templates, and generates electronic medical records by obtaining the corresponding relationship between key information and templates. Obtained from the module library. Specifically, firstly, the characteristic words and/or sentences of the patient's chief complaint question and the characteristic words and/or sentences of the disease question asked by the doctor can be matched with the relevant characteristic words on the medical record template to obtain the characteristic words and/or sentences of the patient's chief complaint question. /or the position of the sentence and the feature word and/or sentence of the disease question; or match the doctor label and patient label with the doctor label or patient label on the medical record template to obtain the corresponding position. Next, the characteristic words and/or sentences about the chief complaint description and the characteristic words and/or sentences for the disease description answered by the patient are placed at corresponding positions, so as to generate an electronic medical record report. It can be seen that the existing way of generating electronic medical records is limited by the electronic template, and each information can only be placed in the corresponding position on the template.

In the embodiment of the present application, instead of using the medical record template provided by the hospital system, the electronic medical record report can be generated directly according to the extracted key information. Specifically, firstly, the doctor label and the patient label, the feature words and/or sentences of the patient's chief complaint question asked by the doctor, and the feature words and/or sentences of the disease question can be vectorized, and the vectorized results can be input into the neural network model processed to generate an image. Then, generate an electronic medical record report by arranging the pictures. In some embodiments, the aforementioned neural network model may be, for example, an open source model dall.e. Based on the open source model dall.e can generate pictures based on the aforementioned key information, and arrange them by medical staff to generate electronic medical record reports. For example, a PPT template can be formed for medical staff to view.

In combination with the above description, it can be seen that the embodiment of the present application divides the voice input of doctors and patients into left and right channels, and sets corresponding role tags respectively, so that the audio data of the left and right channels correspond to the corresponding roles . Further, by identifying the audio data of the left channel and the right channel and extracting key information, the neural network model is used to generate pictures based on the key information, and then the pictures are arranged to generate an electronic medical record report. Based on this, the dialogue text information of doctors and patients can be distinguished, so that the key information extracted by each corresponds to the role label accurately, so as to ensure the generation of correct electronic medical record reports. Furthermore, medical staff can design electronic medical records according to their own needs, without being restricted by the electronic medical record template, so that the generated electronic medical records are diversified.

Fig. 2 is an exemplary diagram illustrating obtaining dialogue text corresponding to a role tag according to an embodiment of the present application. As shown in FIG. 2 , during the consultation process, the doctor and the patient input audio data through the left audio channel 201 and the right audio channel 202 respectively. For example, the doctor can ask the patient through the left sound channel 201, and the patient can answer through the right sound channel 202 or the doctor can ask the patient through the right sound channel 202, and the patient can answer through the left sound channel 201, so as to output respective audio data. In an application scenario, the aforementioned left channel 201 and right channel 202 are set with role tags. As an example, when the doctor asks questions to the patient through the left channel 201 and the patient answers through the right channel 202, the role label of the left channel 201 is set as doctor, and the role label of the right channel 202 is set as patient. Further, the respective audio data are recognized by using, for example, speech recognition software to obtain the corresponding dialogue text, wherein the corresponding role label is displayed before the dialogue text.

For example, in an exemplary scenario, when the doctor asks the patient's main complaint information (such as age, address, etc.) and disease information (such as eye disease, etc.) through the left channel 201, the corresponding medical question text can be obtained through voice recognition , such as "Doctor: How old is it? Doctor: What symptoms do you have in the eyes? Doctor: When did this symptom start and how long has it lasted? Doctor: Have you taken any medication?". Correspondingly, based on the doctor's question, the patient can obtain the corresponding question answer text through voice recognition after answering the aforementioned main complaint information (such as age, address, etc.) and disease information (such as eye disease, etc.) through the right audio channel 202 , such as "Patient: 29; Patient: There is a foreign body sensation in the eye, and tingling and itching; Patient: It has lasted for a week; Patient: No medication".

As mentioned above, after obtaining the dialogue text corresponding to the role label (including the question text and the answer text), the key information related to the consultation can be extracted based on the dialogue text. In the application scenario, the key information can be extracted through the trained entity annotation model, and the key information can include the characteristic words and/or sentences of the patient's main complaint question asked by the doctor, the characteristics and/or sentences of the disease question, and the patient's answer Describe the characteristic words and/or sentences about the chief complaint and the characteristic words and/or sentences for the disease description. For example, taking the dialogue text in Figure 2 above as an example, key information such as "age", "29", "eye symptoms", "a foreign body sensation in the eye, tingling and itching", and "medication" can be extracted. Further, by inputting the aforementioned key information and role labels into the neural network model, images are generated after being processed by the neural network model, and an electronic medical record report is generated by arranging the images, as shown in Figure 3, for example.

Fig. 3 is an exemplary diagram illustrating generating an electronic medical record report according to an embodiment of the present application. As shown in Fig. 3, at first character label (comprising doctor label and patient label) 301 and key information 302 (comprising the feature word and/or sentence of the chief complaint question and/or sentence and disease question feature and/or sentence, patient's question about patient that doctor asks, patient Vectorization 303 is performed on the characteristic words and/or sentences describing the chief complaint and the characteristic words and/or sentences describing the disease). In an exemplary scenario, for example, the vectorized doctor label is (0000), the patient label is (0001), and the vectorized foreign body sensation in the eyes is (00010100). Wherein, the setting of the aforementioned 0 or 1 is determined based on the specific position in the vector where the aforementioned doctor tag, patient tag, and key information are stored. When the received information is included in the vector, the specified position is set to 1, otherwise it is set to 0 . Next, the aforementioned vectorized results are input into the neural network model (eg dall.e) 304, the neural network model is a fully connected structure, including at least one hidden layer, and the output of the neural network model is a plurality of pictures. Among them, each picture contains corresponding information, such as "patient's age", "29", "patient's symptoms", "a foreign body sensation in the eye, tingling and itching, which lasted for a week, and no medication" and "doctor Suggestions", "It is recommended that patients close their eyes and rest more, and look less at electronic devices", etc. Further, the electronic medical record 305 can be generated by arranging the aforementioned multiple pictures as required.

In some embodiments, the neural network model of the embodiment of the present application can also directly input the dialogue text and role label corresponding to the role label, and the neural network model extracts the key information related to the consultation, and the role label and key information Processed into pictures to generate electronic medical records.

Fig. 4 is an exemplary schematic diagram illustrating a generated electronic medical record report according to an embodiment of the present application. As shown in FIG. 4 , after the vectorized character labels and key information are processed through the above neural network model, multiple pictures can be generated, which contain corresponding information. For example, what is shown in the figure includes "patient's age", "29", "patient's symptom description", "foreign body sensation in the eye, tingling, itching, lasted for a week, no medication" and "doctor's advice", " Advise patients to close their eyes and rest more, and look less at electronic equipment" pictures, which are arranged to generate the electronic medical records shown in the picture.

In an implementation scenario, the doctor and the patient may use the wrong voice input device. For example, the doctor corresponds to the patient’s voice channel, and the patient corresponds to the doctor’s voice channel. information, while the patient corresponds to the text information of the doctor, which leads to errors in the electronic medical record report, as shown in Figure 5.

Fig. 5 is an exemplary schematic diagram illustrating generating an erroneous electronic medical record report according to an embodiment of the present application. Figure 5 (a) shows the dialogue text between the doctor and the patient. In this scenario, the text information of the doctor and the patient is inconsistent due to the wrong corresponding vocal channels of the doctor and the patient. For example, "Patient: How old is it? Patient: What symptoms do you have in the eyes? Patient: When did the symptoms start and how long has it lasted? Patient: Have you taken any medication?" and "Doctor: 29; Doctor: There is a foreign body sensation in the eyes, And tingling, itching; doctor: lasted for a week; doctor: no medication". In this scenario, when key information is extracted from the dialogue text, there will be inconsistencies between the role label and the corresponding key information, resulting in electronic medical record reports Incorrect. For example, as shown in (b) in Figure 5, "patient age" is displayed as the doctor's age, "patient's condition" is displayed as "doctor's condition" and "doctor's suggestion" is displayed as "patient's suggestion".

In view of this, the embodiment of the present application proposes to perform speech recognition on the audio data related to the consultation, so as to obtain the respective voiceprint information of the left channel and the right channel, and determine the left channel and the right channel according to the voiceprint information. Is it consistent with the respective role tags. In one embodiment, the respective voiceprint information can also be identified by, for example, voice recognition software or a voice recognition program, and then the identified voiceprint information is compared with the recorded voice of the doctor when registering, thereby determining the left channel and Whether the audio data for the right channel comes from a doctor or a patient determines whether the left and right channels correspond to their respective role labels. When it is determined after comparison that the left and right channels are inconsistent with their respective role labels, you can prompt the doctor and patient to exchange their respective channels and re-enter; or you can also modify the role labels corresponding to the left and right channels . For example, modify the initially set doctor label to a patient label, and change the initially set patient label to a doctor label, so that the left and right channels are consistent with their respective role labels, and the accuracy of generating the electronic medical record template is improved.

In an embodiment, the embodiment of the present application may also provide question prompt information and answer prompt information for the character according to the key information in the character label and the corresponding dialogue text. That is, during the consultation process, the question to be asked by the doctor and the direction to be answered by the patient are prompted, so as to improve the efficiency of the consultation and ensure the comprehensiveness of the consultation. In the implementation scenario, first of all, according to the key information in the corresponding dialogue text, the corresponding value can be calculated by using the knowledge map related to the consultation, and then in response to the corresponding value hitting the path of the knowledge map, provide question prompt information for the corresponding role and answer prompts. It can be understood that the medical knowledge map in the embodiment of the present application is mainly aimed at ophthalmology. It is determined by ophthalmology experts based on the different problems encountered by patients and the corresponding judgments. Confirm the path of the situation, and then determine the knowledge map. Among them, the value calculated above is an index value, through which the knowledge map is searched, and when the index value has a corresponding path in the knowledge map, the doctor or patient is provided with a question prompt or an answer prompt based on the path, for example Figure 6 shows.

Fig. 6 is an exemplary diagram illustrating providing question prompt information and answer prompt information according to an embodiment of the present application. As shown in Figure 6, during the consultation process, when the doctor asks the patient "what are the symptoms of the eyes", the key information will first be extracted from the text information, such as "eye symptoms". Then, based on the eye symptoms, it is calculated through the knowledge map and the corresponding value is obtained. According to the corresponding value, the corresponding path can be found in the knowledge map, and the patient can be provided with answer prompt information. For example, before the patient answers, the patient will be prompted to answer from specific symptoms, onset time, etc., without the need for the doctor to ask questions. Similarly, when the patient has finished answering, the doctor can be prompted for the content of the next question or to give advice without the patient asking further questions. Thus, the efficiency of consultation can be improved, and comprehensive consultation can be ensured.

Fig. 7 is a block diagram showing an exemplary structure of a device 700 for generating an electronic medical record according to an embodiment of the present application. It can be understood that the device implementing the solution of the present application may be a single device (such as a computing device) or a multifunctional device including various peripheral devices.

As shown in FIG. 7, the device of the present application may include a central processing unit or central processing unit (“CPU”) 711, which may be a general-purpose CPU, a dedicated CPU, or other execution units for information processing and program execution. Further, the device 700 may also include a mass storage 712 and a read-only memory ("ROM") 713, wherein the mass storage 712 may be configured to store various types of data, including various audio data, algorithm data, intermediate results, and run Various programs required by the device 700. The ROM 713 can be configured to store data and instructions required for power-on self-test of the device 700, initialization of various functional modules in the system, basic input/output drivers of the system, and booting of the operating system.

Optionally, device 700 may also include other hardware platforms or components, such as shown tensor processing unit ("TPU") 714, graphics processing unit ("GPU") 715, field programmable gate array ("FPGA") ) 716 and Machine Learning Unit ("MLU") 717. It can be understood that although various hardware platforms or components are shown in the device 700, these are only exemplary and non-limiting, and those skilled in the art can add or remove corresponding hardware according to actual needs. For example, the device 700 may only include a CPU, a related storage device and an interface device to implement the method for generating an electronic medical record of the present application.

In some embodiments, in order to facilitate data transmission and interaction with external networks, the device 700 of the present application further includes a communication interface 718, so that it can be connected to a local area network/wireless local area network ("LAN/WLAN") 705 through the communication interface 718, In turn, a connection to a local server 706 or to the Internet ("Internet") 707 can be made via the LAN/WLAN. Alternatively or additionally, the device 700 of the present application may also be directly connected to the Internet or a cellular network based on a wireless communication technology, such as a 3rd generation (“3G”), 4th generation (“4G”) or 4th generation 5th generation (“5G”) wireless communication technology. In some application scenarios, the device 700 of the present application can also access the server 708 and database 709 of the external network as needed, so as to obtain various known algorithms, data and modules, and can store various data remotely, such as for Various types of data or instructions such as audio data, dialogue text, key information, etc. are presented.

The peripheral equipment of the device 700 may include a display device 702 , an input device 703 and a data transmission interface 704 . In one embodiment, the display device 702 may include, for example, one or more speakers and/or one or more visual displays, which are configured to perform voice prompts and/or image and video display for the generated electronic medical records of the present application. The input device 703 may include other input buttons or controls, such as a keyboard, mouse, microphone, gesture capture camera, etc., configured to receive input of audio data and/or user instructions. Data transfer interface 704 may include, for example, serial, parallel, or Universal Serial Bus ("USB"), Small Computer System Interface ("SCSI"), Serial ATA, FireWire ("FireWire"), PCIExpress, and High Definition Multimedia Interface (“HDMI”), etc., configured for data transmission and interaction with other devices or systems. According to the solution of the present application, the data transmission interface 704 can receive the audio data related to the consultation obtained from the left channel and the right channel, and transmit the audio data or various other types of data or results to the device 700 .

The CPU 711, mass memory 712, ROM 713, TPU 714, GPU 715, FPGA 716, MLU 717, and communication interface 718 of the device 700 of the present application can be connected to each other through a bus 719, and data interaction with peripheral devices can be realized through this bus . In one embodiment, through the bus 719, the CPU 711 can control other hardware components in the device 700 and its peripherals.

The device for generating electronic medical records that can be used to execute the present application has been described above with reference to FIG. 7 . It should be understood that the device structure or architecture here is only exemplary, and the implementation manner and implementation entity of the present application are not limited thereto, but changes can be made without departing from the spirit of the present application.

According to the above description in conjunction with the accompanying drawings, those skilled in the art can also understand that the embodiments of the present application can also be implemented by software programs. Therefore, the present application also provides a computer program product. The computer program product can be used to implement the method for generating an electronic medical record described in this application in conjunction with Figures 1 to 6 .

It should be noted that, while operations of the methods of the present application are depicted in the figures in a particular order, this does not require or imply that those operations must be performed in that particular order, or that all illustrated operations must be performed to achieve desirable results. Conversely, the steps depicted in the flowcharts may be performed in an altered order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and/or one step may be decomposed into multiple steps for execution.

It should be understood that the terms "comprising" and "comprising" used in the specification and claims of this application indicate the presence of described features, integers, steps, operations, elements and/or components, but do not exclude one or more other Presence or addition of features, wholes, steps, operations, elements, components and/or collections thereof.

It should also be understood that the terminology used in the description of the present application is only for the purpose of describing specific embodiments, and is not intended to limit the present application. As used in the specification and claims of this application, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly dictates otherwise. It should be further understood that the term "and/or" used in the specification and claims of the present application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

Although the embodiments of the present application are as above, the content described is only an embodiment adopted for the convenience of understanding the present application, and is not intended to limit the scope and application scenarios of the present application. Anyone skilled in the technical field described in this application can make any modifications and changes in the form and details of implementation without departing from the spirit and scope disclosed in this application, but the patent protection scope of this application , must still be subject to the scope defined by the appended claims.

Claims (11)

1. A method for generating an electronic medical record, comprising: Acquiring audio data related to the consultation from the left channel and the right channel, wherein the left channel and the right channel are respectively set with respective role labels; performing speech recognition on the audio data related to the consultation, so as to at least obtain dialogue text corresponding to the role label; Extract key information related to the consultation based on the dialogue text; and Use the neural network model to generate electronic medical records based on key information in the role labels and corresponding dialogue texts. 2. The method of claim 1, further comprising: in response to the character's stereo input, setting the character's stereo input to a stereo format; and Channel splitting is performed on the stereo format to determine the left channel and the right channel. 3. The method according to claim 1, wherein the role tags include a doctor tag and a patient tag, and the dialogue text includes a medical question text corresponding to the doctor tag and an answer to the patient tag text. 4. The method according to claim 3, wherein extracting its key information relevant to medical inquiry based on dialogue text comprises: marking the text of the medical question and the text of the answer; and Key information is extracted from the marked text of the question and the answer. 5. The method according to claim 4, wherein the key information includes at least the characteristic words and/or sentences of the patient's chief complaint question and the characteristic words and/or sentences of the disease problem and the characteristic words and/or sentences of the patient's answer about the chief complaint described by the doctor's question And/or sentences and disease description feature words and/or sentences. 6. The method according to claim 5, wherein using the neural network model to generate the electronic medical record according to the key information in the role label and the corresponding dialogue text comprises: Key information in the character label and the corresponding dialogue text is input into the neural network model to obtain a picture related to the medical record; and An electronic medical record is generated based on the picture. 7. The method of claim 1, further comprising: performing voice recognition on the audio data related to the medical inquiry, so as to obtain the respective voiceprint information of the left channel and the right channel; and Determine whether the left sound channel and the right sound channel are consistent with their respective role labels according to the voiceprint information. 8. The method of claim 1, further comprising: According to the key information in the character label and the corresponding dialogue text, the character is provided with question prompt information and answer prompt information. 9. The method according to claim 8, wherein providing the character with question prompt information and answer prompt information according to the key information in the corresponding dialogue text comprises: According to the key information in the corresponding dialogue text, use the knowledge map related to the consultation to calculate its corresponding value; and In response to the corresponding numerical value hitting the path of the knowledge graph, the corresponding role is provided with question prompt information and answer prompt information. 10. A device for generating electronic medical records, comprising: processor; and The memory stores program instructions for generating electronic medical records, and when the program instructions are executed by the processor, the device implements the method according to any one of claims 1-9. 11. A computer-readable storage medium, on which computer-readable instructions for generating an electronic medical record are stored, and when the computer-readable instructions are executed by one or more processors, the implementation of any one of claims 1-9 Methods.

Images