CN110223685A - New hospital based on DSP speech recognition is registered guidance system - Google Patents

Abstract

It registers guidance system the invention discloses a kind of new hospital based on DSP speech recognition, terminal of registering includes the controller of dsp chip, what is be connected with the controller is used to acquire the voice module for personnel's voice signal of registering, be connected with the controller for showing that relevant information and input or selection are registered the touching liquid-crystal display screen of information, what is be connected with the controller is used to read the RFID card reader module for personnel identity information of registering, for issuing the loudspeaker module of speech prompt information under the voice register modes being connected with the controller, the WiFi module for being communicated with server terminal being connected with the controller；Server includes the PC machine and server local database for being connected into local area network；In the local data base of server end, by each department by number storage, each department's number is lower to index corresponding symptom keyword data structure；The present invention has the characteristics that facilitation, multifarious, keep doctors and patients' system interconnected with it is humanized.

Description

A New Hospital Registration Guidance System Based on DSP Speech Recognition

technical field

The invention relates to a voice registration guidance system, in particular to a novel hospital registration guidance system based on DSP voice recognition.

Background technique

The existing registration methods in the hospital are mainly divided into two methods: manual registration and self-registration machine. The former method uses medical workers to register in designated departments, or medical workers recommend based on experience after listening to the patient's description of their own symptoms. The registration of the department requires the judgment and operation of the medical workers themselves, and in most hospitals, the registration and payment windows are combined into one, and the patient wastes the time of other queuing and payment personnel in the process of describing the condition. The same medical workers need to switch work modes frequently: either registering or paying fees, which brings inconvenience to medical workers and wastes resources; in the latter way, self-registration requires patients to understand their own conditions on the basis of , which can speed up the efficiency of self-selected department registration, but the vast majority of patients will ask experienced medical workers to judge their symptoms and recommend relevant departments for treatment, which has the same shortcomings as the former. Therefore, compared with the present invention, the existing registration method or registration guidance system has the following deficiencies:

1. The registration and payment windows are combined into one, and medical workers need to switch work modes and increase the workload;

2. Occupying the working time of medical workers to listen to the description of the disease, wasting time and resources;

3. Most of the self-registration machines are unmanned, wasting social and public resources;

4. The number of people that the autonomous registration machine can deal with is small, and the interconnection and openness are small;

5. The connectivity and portability are low. A system designed only for a single requirement is difficult to interact with other systems.

SUMMARY OF THE INVENTION

1. The purpose of the invention.

In order to overcome the above problems, the present invention proposes a new hospital registration guidance system based on DSP speech recognition.

2. The technical solution adopted in the present invention.

The invention discloses a novel hospital registration guidance system based on DSP voice recognition, comprising two parts: a registration terminal and a server terminal;

The registration terminal includes a controller of a DSP chip, a voice module connected to the controller for collecting voice signals of registered personnel, and a touch LCD screen connected to the controller for displaying relevant information and inputting or selecting registration information , the RFID card reader module connected with the controller for reading the identity information of registered personnel, the speaker module connected with the controller for issuing voice prompt information in the voice registration mode, connected with the controller The WiFi module used to communicate with the server terminal;

The server includes a PC connected to the local area network and the server local database; in the local database on the server side, each department is stored by number, and the index under each department number corresponds to the symptom keyword data structure;

The registration terminal includes three modules:

In the traditional registration module, jump to the department selection interface, independently select the corresponding department and finally click OK on the interface to complete the registration in this way;

Voice registration module, the registration terminal sends out a prompt tone, and jumps to the voice recognition interface at the same time, the voice module recognizes the voice description and converts it into text and displays it on the interface. Match the keywords in the interface, and finally display the best recommended department on the interface. After the patient confirms the choice, the registration in this way is completed;

In the text registration module, the interface jumps to the keyword selection interface, where the patient selects words related to their own conditions. If the keywords on the interface cannot describe the most obvious symptoms of the patient, the patient can click on other words at the bottom of the interface. Symptom description, enter the information of the symptom by typing; after all the patient's symptom information is selected or described, click OK on the interface, and the background server will match the keywords in the database with the keywords in the patient's description of their symptoms. , and finally the best recommended department will be displayed on the interface, and the patient will complete the registration in this way after confirming the choice. During a period of time, if a certain disease keyword is frequently entered by the patient, the server terminal software will add the keyword into the database for future matching;

The server terminal is in the same local area network as each registration terminal, and monitors the registration terminal connection and sending instructions in real time. The server terminal displays a plan map of the hospital, and marks the location of each registration terminal on the map. If the server establishes a communication connection, it will be marked on the map that the registered terminal is faulty;

The traditional registration module returns the number of the corresponding department, which is parsed and displayed by the microcontroller. After confirmation, the server terminal transmits the patient-related identity information and registration information to the computer of each department; in the voice registration and text registration, the server terminal All keywords need to be processed by the K-nearest neighbor method, and the corresponding results are returned to the registration terminal. After the patient confirms, the server terminal transmits the patient-related identity information and registration information to the computer of each department.

Further, in the traditional registration module, the microcontroller numbers each department. After the patient clicks to select the corresponding department, the microcontroller sends the corresponding numbering information of the department to the server through the WiFi module, and the server matches the corresponding department according to the number. department, and feedback the department information to the registration terminal through the WiFi module to complete the department's reservation and the process of sending the patient information identified earlier to the department.

Further, in the voice registration module, the microcontroller performs analog-to-digital conversion on the voice signal transmitted by the voice module, and compares and recognizes according to the corresponding digital signal-character library, recognizes the narrated text, and displays it on the interface. After confirming the recognition result, the microcontroller receives the confirmation instruction, converts the Chinese characters to Unicode characters, and transmits the character data to the server terminal through the WiFi module. Keyword locking, the locked keywords are matched with the keyword library in the database using the K-nearest neighbor method, and finally the optimal selection is determined according to multiple keywords. Finally, the server feeds back the number of the optimally selected department to the registration through the WiFi module. The terminal completes the reservation of the department and the process of sending the previously identified patient information to the department.

Further, in the text registration module, the microcontroller stores the symptom words selected or described on the interface in the array. After the patient finally confirms the description, the microcontroller receives the confirmation instruction and stores the Chinese characters in the array. Perform Unicode character conversion, and transmit the character data to the server terminal through the WiFi module. The server terminal performs the inverse process of Unicode to Chinese characters, and uses the K-nearest neighbor method to match each keyword in the Chinese character array with the keyword library in the database. Finally, the optimal selection is determined according to multiple keywords. Finally, the server feeds back the number of the optimally selected department to the registration terminal through the WiFi module to complete the reservation of the department and the process of sending the previously identified patient information to the department.

3. The technical effect produced by the present invention.

This system meets the needs of a variety of registration methods. The traditional registration method is aimed at patients who are familiar with the existing self-registration forms. They have a better understanding of their own conditions. By providing this registration method, the previous users' familiarity with the new registration system is buffered. On the basis of the original registration method, patients are exposed to and familiar with the new registration method; the voice registration method has replaced the medical workers, and the process of describing the characteristics of their symptoms to the medical workers by patients can be replaced by the voice module to describe their own conditions. The process of judging the patient's condition and recommending medical treatment through the worker's own experience can be replaced by matching the background database information; the text registration method is oriented to the process of summarizing the patient's condition and guiding the recommendation registration process through medical workers. Keyword selection completes the guiding process of the original medical staff to the patient. The above-mentioned voice registration and text registration forms relieve the work pressure of medical workers to a certain extent, save the time resources of other patients, and indirectly improve the utilization efficiency of public resources and the efficiency of hospital patients.

Description of drawings

Figure 1 is a structural block diagram of the new hospital registration guidance system.

Figure 2 shows the home page display interface.

Figure 3 shows the department selection interface.

Figure 4 is the voice registration recognition interface.

Figure 5 shows the working process of voice registration.

Figure 6 is a text registration recognition interface.

Figure 7 shows the working process of text registration.

Detailed ways

Example 1

The invention relates to a novel hospital registration guidance system based on DSP speech recognition, the hardware structure of which includes a registration terminal and a server terminal. The registration terminal includes a microcontroller with a DSP chip, a voice module connected to the controller for collecting voice signals of registered personnel, and a touch screen connected to the controller for displaying relevant information and inputting or selecting registration information. A liquid crystal display screen, an RFID card reader module connected to the controller for reading the identity information of registered personnel, a speaker module connected to the controller for issuing voice prompt information in a voice registration mode, and a speaker module connected to the controller A WiFi module connected to the server for communicating with the server terminal; wherein, the server includes a PC connected to the local area network and server software installed on the PC. The server software is developed in C# language, including the server local database. In the present invention, the patient comes to the registration terminal and puts on his identity information card. The registration terminal first determines the patient's identity information by reading the card. The home page of the registration terminal provides three registration methods, namely: traditional registration, voice registration, and text registration. registered. Among them, traditional registration is the existing registration method of the hospital. After the patient clicks this registration method on the home page, the interface jumps to the department selection interface. The patient selects the established department and clicks OK to complete the registration under this method; After clicking the voice registration method on the home page, the registration terminal will issue a voice prompt to remind the patient to describe his symptoms to the voice module. After the server database is matched and other processes, the patient waits for a short time and then the display interface will prompt the department with the most similar symptoms described. The patient clicks OK to register the department to complete the registration in this way; after the patient clicks the text registration method on the home page , the interface jumps to the key symptom selection interface, where different page numbers display common patient symptom information, such as: "running nose", "headache", "stomach pain", "cough", etc. Click on the symptom information that matches your own (you can turn the page to select), if your own special symptom information is difficult to find or not set on the interface, the patient can click on other symptom descriptions at the bottom of the interface, enter the symptom information by typing, all symptom information After selecting or describing, click OK on the interface. After the microcontroller transcodes the disease information, character transmission, and server database matching, etc., the patient waits for a short time and the interface will prompt the most similar disease characteristics. Department, the patient clicks OK to register the department to complete the registration in this way. The above three methods correspond to different registration groups respectively. The registered patients faced by traditional registration know their condition and already know in advance which department they should register. Therefore, the patient only needs to select the department to complete the registration. This method is also the most convenient and relieves the pressure of ordinary queuing for registration or payment windows; voice registration faces the situation that registered patients do not know which department they go to for treatment in advance and can accurately describe their own conditions. Under the existing circumstances, registered patients often describe the condition of medical workers in the queue for registration or payment window. The medical workers listen to the patient's symptom information and complete the patient's registration through their own experience. This method of registration can best alleviate the common queue for registration. Or the pressure of the payment window; the text registration is faced with the situation that the patient does not know in advance which department he goes to see a doctor and needs to use auxiliary text to accurately describe his condition. The difference between this method and the voice registration lies in the degree of description of self-symptom information, and the degree of convenience is between the three methods. The invention has the characteristics of convenience and diversity, and introduces the Internet of Things information technology into the medical system, which is a supplement to the existing doctor-patient system and further interconnects and humanizes the doctor-patient system.

Example 2

The novel hospital registration guidance system designed by the present invention includes two parts: a registration terminal and a server terminal. Before the patient comes to the self-registration machine, first place a card with its own identity information, such as a medical insurance card or social security card. Three registration methods are provided on the display screen of the registration terminal, among which the home page interface of the LCD screen is shown in Figure 2.

On the home page display interface, the user clicks traditional registration, and the screen jumps to the department selection interface. After the patient judges his own condition, click to select the corresponding department and finally click OK on the interface to complete the registration in this way. Among them, the department selection interface is shown in Figure 3.

In this registration method, the microcontroller numbers each department. After the patient clicks to select the corresponding department, the microcontroller sends the corresponding numbering information of the department to the server through the WiFi module, and the server matches the corresponding department according to the number. The information is fed back to the registration terminal through the WiFi module to complete the department's reservation and the process of sending the previously identified patient information to the department.

On the home page display interface, the user clicks on voice registration, the registration terminal will send out a prompt tone of "Please start describing your symptoms", and the screen will jump to the voice recognition interface. The text recognized according to the patient's voice description will be displayed on this interface. After confirming that the description information is correct, click OK on the interface. The background server matches the keywords in the patient's description of their symptoms with the keywords in the database, and finally displays the best recommended department on the interface. registration in this way. The voice registration recognition interface is shown in Figure 4.

In this registration method, the microcontroller converts the voice signal transmitted by the voice module to analog-to-digital conversion, and performs comparison and recognition according to the corresponding digital signal-character library, and recognizes the text described by the patient and displays it on the interface. After the recognition result is finally confirmed, the microcontroller receives the confirmation instruction, converts the Chinese characters to Unicode characters (hexadecimal, etc.), and transmits the character data to the server terminal through the WiFi module, and the server terminal performs the reverse process of Unicode to Chinese characters. Keyword locking is performed on the Chinese paragraph, and the locked keywords are matched with the keyword library in the database using the K-nearest neighbor method, and finally the optimal selection is determined according to multiple keywords. Finally, the server passes the number of the optimal selection department through The WiFi module feeds back to the registration terminal to complete the department's reservation and the process of sending the previously identified patient information to the department. Among them, the working process of voice registration is shown in Figure 5.

On the homepage display interface, the user clicks on the text to register, and the interface jumps to the keyword selection interface, where the patient selects words related to their own conditions (the page can be turned to select), if the keywords on the interface cannot describe the patient's most For obvious symptoms, the patient can click on the description of other symptoms at the bottom of the interface and enter the information of the symptom by typing. After all the patient's symptom information is selected or described, click OK on the interface, the background server matches the keywords in the patient's description of their symptoms with the keywords in the database, and finally displays the best recommended department on the interface. , the patient will complete the registration in this way after confirming the choice. During a period of time, if a certain disease keyword is frequently entered by the patient, the server terminal software will add the keyword into the database for future matching. The text registration recognition interface is shown in Figure 6.

In this registration method, the microcontroller stores the symptom words selected or described by the patient on the interface in an array in the software program design. After the patient finally confirms the description, the microcontroller receives the confirmation instruction and stores the number of Chinese characters The group performs Unicode character conversion (hexadecimal, etc.), and transmits the character data to the server terminal through the WiFi module. The server terminal performs the inverse process of Unicode to Chinese characters, and compares each keyword in the Chinese character array with the keyword database in the database. The K-nearest neighbor method is used for matching, and the optimal selection is finally determined according to multiple keywords. Finally, the server feeds back the number of the optimally selected department to the registration terminal through the WiFi module, completes the reservation of the department and sends the previously identified patient information to the department. the process of. Among them, the working process of text registration is shown in Figure 7.

The server terminal is in the same local area network as each registration terminal, and monitors the registration terminal connection and the request to send instructions in real time. The server terminal host computer software is programmed in C# language. The main page displays the hospital plan map, and each The location of the registered terminal. If a registered terminal has not established a communication connection with the server for a long time, it will be marked on the map that the registered terminal is faulty. In the local database built by the server software, it is numbered and stored according to each department, and the corresponding symptom keywords are corresponding to each department number. After the connection between the registration terminal and the server software is established, in the traditional registration mode, only the number of the corresponding department is returned, which is parsed and displayed by the microcontroller. In both voice registration and text registration, the server terminal needs to perform K-nearest neighbor processing on keywords, and returns the corresponding results to the registration terminal. After the patient confirms, the server terminal transmits the patient's relevant identity information and registration information to each department. computer side.

The above-mentioned embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited by the above-mentioned embodiments, and any other changes, modifications, substitutions, combinations, The simplification should be equivalent replacement manners, which are all included in the protection scope of the present invention.

Claims (4)

1. The guidance system 1. a kind of new hospital based on DSP speech recognition is registered, it is characterised in that: including terminal and the service of registering Device terminal two parts；

   Terminal of registering include dsp chip controller, be connected with the controller be used for acquire personnel's voice signal of registering Voice module, be connected with the controller for showing that the register touch LCD of information of relevant information and input or selection is shown Screen, the RFID card reader module for reading personnel identity information of registering being connected with the controller are connected with the controller Under the voice register modes connect for issue speech prompt information loudspeaker module, be connected with the controller be used for and clothes The WiFi module of device terminal of being engaged in communication；

   Server includes the PC machine and server local database for being connected into local area network；In the local data base of server end, By each department by number storage, each department's number is lower to index corresponding symptom keyword data structure；

   Terminal of registering includes three kinds of modules:

   Traditional registration module, jumps to department's selection interface, independently selects corresponding department and finally clicks the determination on interface Complete registering for this kind of mode；

   Voice registration module, terminal of registering issues prompt tone, while jumping to speech recognition interface, and voice module identification voice is retouched It states and is converted to text displaying on interface, after confirmation description information is correct, background server is according to the pass in illness descriptive statement Keyword, is matched with the keyword in database, after finally showing that the department most preferably recommended, patient determine selection on interface Complete registering for this kind of mode；

   Text registration module, interface jump to Keyword Selection interface, and patient selects relevant to itself state of an illness on the surface Word, if the keyword on interface can not describe the most apparent symptom of patient itself, patient can click interface bottom Other symptoms description, pass through the information of the typewriting input symptom；After patient's symptomatology information has been selected or described, click Determination on interface, background server according to patient to the keyword in itself illness descriptive statement, with the key in database Word is matched, and is finally shown on interface and is completed registering for this kind of mode after the department most preferably recommended, patient determine selection.? In a period of time, if some illness keyword, by the customized input of patient's high-frequency, server terminal software can be by the key Word is added into the matching carried out in database in the future；

   Server terminal is in same local area network with each terminal of registering, and monitors register terminal connection and transmission instruction in real time Request, server terminal show hospital's plane map, the position of each terminal of registering is marked on map, if there is some extension Number terminal is not established with server communicate to connect for a long time, then terminal failure of registering is marked on map；

   Traditional registration module returns to the number of corresponding department, is shown after the parsing of microcontroller end, confirmed, server terminal Patient's related identification information and information of registering are transmitted to the computer end of each department；It registers in voice and text is registered mode Under, server terminal is required to carry out k-nearest neighbor processing to keyword, returns to corresponding result to terminal of registering, patient's confirmation Afterwards, patient's related identification information and information of registering are transmitted to the computer end of each department by server terminal.
2. The guidance system 2. the new hospital according to claim 1 based on DSP speech recognition is registered, it is characterised in that: pass Unite registration module, each department is numbered in microcontroller, patient clicks select corresponding department after, microcontroller is by department Corresponding number information is sent to server by WiFi module, and server is matched to corresponding department according to number, by department Information feeds back to terminal of registering by WiFi module, completes the predetermined of department and sends recognized patient early period to the department The process of information.
3. The guidance system 3. the new hospital according to claim 1 based on DSP speech recognition is registered, it is characterised in that: language Sound registration module, the voice signal that microcontroller transmits voice module carry out analog-to-digital conversion, and according to corresponding number Signal-character repertoire compares identification, identifies described text, and be shown on interface, after confirming recognition result, micro-control Device processed receives confirmation instruction, Chinese text is carried out the conversion of Unicode character, and be transmitted to character data by WiFi module Server terminal, server terminal carries out Unicode and turns Chinese text inverse process, and carries out keyword locking to Chinese paragraph, The keyword of locking is matched with the keywords database in database using k-nearest neighbor, is finally determined according to multiple keywords Optimal selection completes department finally, the number of optimal selection department is fed back to terminal of registering by WiFi module by server Process predetermined and that the patient information that early period identifies is sent to the department.
4. The guidance system 4. the new hospital according to claim 1 based on DSP speech recognition is registered, it is characterised in that: text This registration module, microcontroller store symptom word select on interface or description in array, and patient is final After the completion of confirmation description, microcontroller receives confirmation instruction, Chinese text array is carried out the conversion of Unicode character, and pass through Character data is transmitted to server terminal by WiFi module, and server terminal carries out Unicode and turns Chinese text inverse process, and right Each keyword is matched with the keywords database in database using k-nearest neighbor in Chinese text array, finally according to multiple Keyword determines optimal selection, finally, the number of optimal selection department is fed back to terminal of registering by WiFi module by server, It completes the predetermined of department and sends the process for the patient information that early period identifies to the department.