KR101868404B1 - Emergency Dispatch System and Method based on Korean Speech Recognition Technology - Google Patents

Abstract

The present invention relates to an emergency situation control system and a control method for extracting important information and guiding measures such as severity classification, multiple dispatch order, and emergency consultation at the stage of receiving a speech recognition based notification by a general situation room supervisor using Korean speech recognition technology Said system comprising: a speech recognition based report acceptance solution; A multi-criteria severity classification solution in which the severity of patients is classified based on the information received from the report acceptance solution and displayed on the display; A multi-tasking system command solution for instructing the emergency dispatch based on the information received from the report acceptance solution and displaying on the display to provide necessary emergency action to the emergency personnel; And an ambulance counseling solution that displays on the display a guide to the emergency agent before the emergency agent based on information received from the report acceptance solution.

Description

Technical Field [0001] The present invention relates to a speech recognition system,

The present invention relates to an emergency management system and a control method, and more particularly, it relates to an emergency management system and a control method, which extract important information at a stage of receiving a speech recognition based notification by a general situation room supervisor using Korean speech recognition technology and classify the severity, And an emergency management system and control method for guiding the emergency management system.

Fire, disaster, disaster and other emergency situations that require rescue and rescue. Attempting for firefighting activities such as report reception, communication management, and collection and dissemination of disaster information In the general situation room of firefighting headquarters or firefighting department, computer assistant, Respectively. In the case of an emergency, the Suwa personnel perform the reporting of the occurrence of various disasters such as fire, the order of notification, the delivery and dispatch, and carry out the report according to the 'emergency rescue standard system'. However, Consultation manual 'of' introduction part support protocol '. Therefore, the surgeon performs the report reception work according to the 'emergency rescue standard system'. In the case of emergency, the surgeon grasps the outline, the patient's age, the patient's consciousness, the patient's respiration, the patient's cardiac arrest / In particular, the emergency dispatch situation of the outline should be ordered. However, there is a problem that the efficiency of the repository work is deteriorating because the 'emergency rescue standard system' and 'repository protocol', which are essential for the work of Suwa personnel, are not connected.

In addition, in the current Emergency Situation Supervisor, the assistant supervisor enters the basic information into the system in the process of responding to the emergency report telephone call, decides whether or not to send an ambulance, and then the situation supervisor collects additional information about the patient, And first aid instruction. If the emergency control system can not be operated smoothly during this process, it may be fatal to the emergency patients if delay of the ambulance dispatch time and the lack of prompt emergency measures are not provided.

Therefore, it is necessary to develop technology to effectively support the work of Subo personnel and to increase the speed and accuracy of responding to crisis situations.

Korean Patent No. 10-0759916 relates to an emergency medical information system for selecting a severity classifying and transfer hospital. In the emergency situation before the hospital, information on the emergency activity log created at the time of emergency rescue is inputted through a web application Based on this, we describe a technology related to the emergency medical information system that determines the severity classification of the patient and selects the optimal transfer hospital of the patient accordingly. However, since the above technique can be applied in the situation where the emergency activity log is written, there is a problem that it is far from the urgency of the crisis response.

Korean Patent Publication No. 10-0759916

The present invention has been developed in order to solve the above problems, and it is an object of the present invention to apply the Korean speech recognition technology to the emergency situation control system, to support the promptness and efficiency in receiving the 119 report of the subordinate agent, We suggest a system and method for supporting emergency measures.

The system comprises an automatic input unit for automatically inputting basic data by using Korean speech recognition, and an automatic input unit for automatically inputting basic data from the basic data, A semantic extraction unit for extracting semantics, and an automatic classification unit for automatically classifying an emergency situation based on the extracted semantics to provide a multi-criteria severity classification solution, a multi-mission system command solution, and an emergency response consulting solution. ; A multi-criteria severity classification solution for sorting and displaying the severity of patients based on the information received from the report acceptance solution; A multi-tasking system command solution for instructing the emergency dispatch based on the information received from the report acceptance solution and displaying on the display to provide necessary emergency action to the emergency personnel; And an emergency rescue consulting solution for displaying on the display a guidance for emergency action before arrival of the emergency agent on the basis of the information received from the report acceptance solution, wherein the information extracting unit of the voice recognition based report accepting solution comprises: The DB includes a voice recognition notification information DB, a high-risk patient DB, an advanced protocol guideline DB, a dispatching transfer DB , And a telephone reception DB. Emergency situation control system using Korean speech recognition technology:

[Formula 1]

는 두 문자열 중 a 문자열의 i번째 문자가 b 문자열의 j번째 문자와 같을 경우 (

), 값은 0과 같고, 다르면 값이 2와 같음을 의미함)을 제공한다. (In the above formula

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, and if it is different, the value is equal to 2).

The present invention also relates to a method of treating a patient suffering from a condition selected from the group consisting of headache, chest pain, abdominal pain, back pain, labor pain, other pain, unconsciousness, airway water, dyspnea, respiratory arrest, palpitations, 31 of 31 cases of syncope, nausea or vomiting, diarrhea or constipation, dysuria, hemoptysis, hematochezia, hemorrhage, vaginal bleeding, other bleeding, hyperthermia, hypothermia, dizziness or dizziness, paralysis, general weakness, mental disorders, Which is classified into a plurality of categories.

The present invention is also characterized in that said emergency dispatch comprises a report consisting of hanging, strangulation, suffocation and drowning; And a patient state composed of whether or not to breathe and consciousness, using the Korean speech recognition technology.

The present invention is also characterized in that the personal information includes information on the identity of the applicant, gender, age under 4 years, age 5 to 14, age 15 to 39, age 40 to 69, , And an emergency situation control system using Korean speech recognition technology.

According to the present invention, the basic data automatic input unit of the voice recognition based report acceptance solution is a voice recognizer composed of an acoustic model, a language model, and a pronunciation model, and the acoustic model includes a triphone (Triphone: The language model is a language model in which the words that form the sentence by dividing the word are divided into 1-gram (Unigram), 2-gram (Bigram) grouped by two consecutive 3-grams (Trigram) A 4-gram grouped by 4, and a 5-gram grouped by 5 consecutive, and the pronunciation model provides an emergency situation control system using a Korean speech recognition technology, which is generated by a predetermined sounding sequence generator.

The present invention also provides a method of controlling an emergency situation using a Korean speech recognition technology having a computer solution including a DB, wherein the method comprises an automatic input adding acoustic model of a speech recognition based notification accepting solution provided in the emergency situation control system, A language model, and a pronunciation model; Extracting a meaning including a plurality of categories of the patient's main symptom, emergency call, and personal information using the Lebenstein distance measurement model in which the information extracting unit of the report acceptance solution is defined in the basic data as shown in Equation 1 below; Providing the emergency classification information obtained from the extracted meaning to the multi-criteria severity classification solution, the multiple dispatch system command solution, and the emergency response consulting solution provided in the emergency management system; Classifying the degree of severity of the patient based on the semantic information received from the report acceptance solution and displaying it on the display; Displaying on the display the multi-dispatch system command solution instructing the emergency dispatch based on the information received from the notification acceptance solution and providing the necessary emergency response to the first-aid agent; And displaying on the display the first aid consultation solution to guide a first aid action before the emergency agent on the basis of information received from the notification acceptance solution, wherein the acoustic model comprises a triphone (Triphone) The language model is a model in which words constituting a sentence are divided into 1-gram (Unigram), 2-gram (Bigram) grouped by two consecutive words, 3-gram (Trigram) Wherein the speech model is generated by a predetermined sounding sequence generator, and the DB includes a speech recognition report information DB, a high-risk patient DB, an advanced protocol Emergency Situation Control Method Using Korean Speech Recognition Technology, Including Guideline DB, Departure Transfer DB, and Telephone Reception DB:

[Formula 1]

는 두 문자열 중 a 문자열의 i번째 문자가 b 문자열의 j번째 문자와 같을 경우 (

), 값은 0과 같고, 다르면 값이 2와 같음을 의미함)을 제공한다. (In the above formula

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, and if it is different, the value is equal to 2).

The present invention also relates to a method of treating a patient suffering from a condition selected from the group consisting of headache, chest pain, abdominal pain, back pain, labor pain, other pain, unconsciousness, airway water, dyspnea, respiratory arrest, palpitations, 31 of 31 cases of syncope, nausea or vomiting, diarrhea or constipation, dysuria, hemoptysis, hematochezia, hemorrhage, vaginal bleeding, other bleeding, hyperthermia, hypothermia, dizziness or dizziness, paralysis, general weakness, mental disorders, And the emergency dispatch includes a report consisting of a neck, a strangulation, a suffocation, and a drowning in water; And a patient state consisting of whether or not breathing and consciousness are made. The personal information includes the identity of the applicant, gender, age under 4 years, age 5 to 14, age 15 to 39, age 40 to 69 And an age of 70 years or more, which is divided into a plurality of age groups.

The system and method for emergency situation control using the Korean speech recognition technology of the present invention can effectively support the work of the backup staff by applying the speech recognition technology to the emergency situation control and improve the promptness and accuracy of response to the crisis situation, The development of the emergency control system used has the advantage of improving the current situation in which the training of the situation room personnel is insufficient.

1 is a block diagram of an emergency situation control system using a Korean speech recognition technology according to an embodiment of the present invention.
FIG. 2 is a conceptual diagram illustrating a solution and DB of an emergency situation control system using Korean speech recognition technology according to an embodiment of the present invention.
3 is a conceptual diagram illustrating a display of an emergency situation control system using a Korean speech recognition technology according to an embodiment of the present invention.
FIG. 4 is a conceptual diagram schematically illustrating a Lebenstein distance-based semantic extraction process according to an embodiment of the present invention.
FIG. 5 is a conceptual diagram illustrating a display of a multi-dispatch system command solution in an emergency management system using Korean speech recognition technology according to an embodiment of the present invention.
Figure 6 shows a flowchart of an emergency dispatch scenario, in accordance with one embodiment of the present invention.

Various aspects are disclosed with reference to the drawings. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more aspects. It will be appreciated, however, that such aspects may be practiced without the specific details of each. The following description and the annexed drawings set forth in detail certain illustrative aspects of one or more aspects. However, it is to be understood that such aspects are illustrative and that some of the various ways may be used in various aspects of the principles, and that the description set forth is intended to include all such aspects and their equivalents.

Various aspects and features will be presented by a system that may include multiple devices, modules, and the like. It should also be understood and appreciated that the various systems may include additional devices, components, components, etc., and / or may not include all of the devices, components, components, etc. discussed in connection with the drawings.

As used herein, the terms " an embodiment, "" an embodiment, "" an example, "etc. should not be construed as advantageous or advantageous over other aspects or designs. The terms "system", "solution", "display" and the like used herein generally mean a computer-related entity, for example, hardware, a combination of hardware and software, .

In addition, the term "or" is intended to mean " exclusive or " That is, it is intended to mean one of the natural inclusive substitutions "X uses A or B ", unless otherwise specified or unclear in context. That is, X uses A; X uses B; Or when X uses both A and B, "X uses A or B" can be applied in any of the above cases. It is also to be understood that the term "and / or" as used herein refers to and includes all possible combinations of one or more of the listed items.

Also, the term " comprises "and / or" comprising "means that there is a corresponding feature, step, operation, module, and / Elements, and / or the presence or addition of such groups. In addition, although the terms first and second in the present specification can be used to describe various elements, these elements are not limited by these terms. That is, these terms are only used to distinguish between two or more components, and should not be construed as meaning order or priority. Also, unless the context clearly dictates otherwise or to the contrary, the singular forms in this specification and claims should generally be construed to mean "one or more. &Quot;

FIG. 1 is a block diagram of an emergency situation control system using a Korean speech recognition technology according to an embodiment of the present invention. FIG. 2 is a block diagram of an emergency situation control system using a Korean speech recognition technology, according to an embodiment of the present invention. Solution and DB. 3 is a conceptual diagram illustrating a display of an emergency situation control system using a Korean speech recognition technology according to an embodiment of the present invention. The system according to an embodiment of the present invention is an emergency situation control system using Korean speech recognition technology having a computer solution including a display and a DB. The system according to an embodiment of the present invention includes a Korean speech recognition based report acceptance solution; A multi-criteria severity classification solution in which the severity of patients is classified based on the information received from the report acceptance solution and displayed on the display; A multi-tasking system command solution for instructing the emergency dispatch based on the information received from the report acceptance solution and displaying on the display to provide necessary emergency action to the emergency personnel; And an ambulance counseling solution that displays on the display a guide to the emergency agent before the emergency agent based on information received from the report acceptance solution.

In one embodiment of the present invention, the Korean speech recognition based report acceptance solution includes an automatic input unit for automatically inputting basic data using Korean speech recognition, a semantic extraction unit for extracting semantics from the basic data, An automatic classification section for automatically classifying the emergency situation into a multi-criteria severity classification solution, a multi-mission system command solution, and an emergency classification consulting solution.

The main symptom of the multiple categories of patients according to an embodiment of the present invention may include headache, chest pain, abdominal pain, back pain, labor pain, other pain, conscious disorder, airway water, dyspnea, respiratory arrest, palpitations, , Seizures, syncope, nausea or vomiting, diarrhea or constipation, dysuria, hemoptysis, hematochezia, hemorrhage, vaginal bleeding, other bleeding, hyperthermia, hypothermia, dizziness or dizziness, paralysis, general weakness, mental disorders, Fractures can be divided into 31 categories. In addition, the emergency dispatch according to an embodiment of the present invention may include a report consisting of a neck, a strangulation, a suffocation, and a drowning in water; And patient state consisting of whether or not breathing and consciousness. In addition, the personal information according to an embodiment of the present invention may be provided to a person who is notified of the identity of the applicant, sex, and a person under age 4, a person aged 5 to 14, a person aged 15 to 39, a person aged 40 to 69, It is possible to divide by the age divided into sections.

In one embodiment of the present invention, the basic data automatic input unit of the speech recognition based report acceptance solution is a speech recognizer composed of an acoustic model, a language model, and a pronunciation model. The acoustic model includes a triphone The language model is a model in which a word is divided into a plurality of words and sentence words are divided into individual words of 1-gram (Unigram), two consecutive 2-grams (Bigram), 3 consecutive 3-grams ), A 4-gram group consisting of 4 consecutive 4-grams, and a 5-gram group consisting of 5 consecutive 5-grams, and the pronunciation model is generated by a predetermined sounding sequence generator.

In one embodiment of the present invention, the information extracting unit of the voice recognition-based report accepting solution extracts a plurality of categories of patient main symptoms, emergency calls, and personal information using the Lebensch's distance measurement model defined as Equation 1 below And the DB includes a voice recognition notification information DB, a high-risk patient DB, an advanced protocol guideline DB, a mission transfer DB, and a telephone reception DB.

[Formula 1]

는 두 문자열 중 a 문자열의 i번째 문자가 b 문자열의 j번째 문자와 같을 경우 (

), 값은 0과 같고, 다르면 값이 2와 같음을 의미한다. 일반적인 레벤슈타인 거리 측정에서의 패널티 값은 1을 보편적으로 사용하지만 본 발명에서는 문자열이 달라질 경우 거리 값이 더 커지도록 2를 선택한다.In the above formula

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, otherwise the value is equal to 2. The penalty value in the general Levenshtein distance measurement is generally 1, but in the present invention, 2 is selected so that the distance value becomes larger when the string is changed.

Using the Lebenstein distance measurement between the two strings a and b, we can obtain the Lebenstein distance ratio as shown in Equation 2 below.

[Formula 2]

Here, | a | and | b | are the lengths of two strings. The obtained ratio of Lebensch 's distance is used for extracting meaning by comparing with threshold value of core recognition target. For example, as shown below, the threshold value is set according to the length of the word to be extracted. For example, in the case of word extraction in category such as sex, age, report history, If the extracted words are not exactly the same, do not extract them. Otherwise, set them to 0.8 to 0.88 when the extracted words are long.

FIG. 4 is a conceptual diagram schematically illustrating a Lebenstein distance-based semantic extraction process according to an embodiment of the present invention. As shown in FIG. 4, the Lebenstein distance-based semantic extraction algorithm extracts the key recognition objects by comparing and analyzing the templates of the key recognition target and the speech recognition result sentences of the speech source of the speech room. In one embodiment of the invention, the core subject is designed to extract 31 main symptoms (Table 1), emergency response (Table 2) and personal details (Table 3) extracted in one embodiment of the present invention .

[Table 1] Symptoms of 31 categories of patients

[Table 2] Emergency dispatch

[Table 3] Personal Information

In one embodiment of the present invention, a template for each object can be configured for extracting a key recognition target. Templates representing / expressing the 31 main patient symptoms can be constructed as shown in Table 4 based on the symptomatic language expression of the Korean symptom expression study in the emergency notification telephone.

[Table 4] Example of Korean symptom expression in emergency call report

If the language expression of any symptom is included in the manifestation of several main symptoms, if the main symptom can be detected, the surgeon looks at the main symptom displayed and reduces the main symptom object through the question. It is a method for optimizing the 31 symptom expression templates on the scale of the transcription corpus that has been investigated so far. In an embodiment of the present invention, the emergency dispatch template configuration in the emergency call center forms a template based on the keyword. Key words are words that indicate a critical situation and should not be able to identify other main symptoms or personal information. It includes breathing / consciousness that can be commonly included in all 31 symptomatic expressions in an emergency response. Included in the emergency response is the absence of breathing or unconsciousness because it can be a very critical condition, such as a cardiac arrest.

FIG. 5 is a conceptual diagram illustrating a display of a multi-dispatch system command solution in an emergency management system using Korean speech recognition technology according to an embodiment of the present invention. In one embodiment of the present invention, the template for personal information is simple personal information and personal identity. The reason why the applicant is the most important person in the reception of an emergency report is because, if the person himself / herself does the report reception, it is judged that there is breathing and consciousness, and the probability of emergency is lowered. We construct a template to extract two key information from a single word for words that can predict sex or age.

The present invention relates to an emergency situation control method using a Korean speech recognition technology, the method comprising: an automatic input adding acoustic model of a speech recognition based notification accepting solution provided in an emergency situation control system having a computer solution including a DB; Automatically inputting basic data with a speech recognizer composed of speech models; Extracting a meaning including a plurality of categories of the patient's main symptom, emergency call, and personal information using the Lebenstein distance measurement model in which the information extracting unit of the report acceptance solution is defined in the basic data as shown in Equation 1 below; Providing the emergency classification information obtained from the extracted meaning to the multi-criteria severity classification solution, the multiple dispatch system command solution, and the emergency response consulting solution provided in the emergency management system; Classifying the degree of severity of the patient based on the information provided by the multi-criteria severity classification solution from the report acceptance solution and displaying on the display; Displaying on the display the multi-dispatch system command solution instructing the emergency dispatch based on the information provided from the report acceptance solution and providing the necessary emergency response to the emergency personnel; And displaying on the display the first aid consultation solution to guide the first aid action before the first aid agent based on the information provided from the report acceptance solution.

[Formula 1]

는 두 문자열 중 a 문자열의 i번째 문자가 b 문자열의 j번째 문자와 같을 경우 (

), 값은 0과 같고, 다르면 값이 2와 같음을 의미한다.In the above formula

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, otherwise the value is equal to 2.

According to an embodiment of the present invention, the acoustic model is a context-dependent triphone (Triphone: Tri Phoneme) model, and the language model includes words that form sentences by dividing a word into 1-gram (Unigram) (Bigram), two 3-grams (Trigram), three 4-grams (4-grams), and five 5-grams (5-grams) And the DB includes a voice recognition report information DB, a high-risk patient DB, an advanced protocol guideline DB, a mission transfer DB, and a telephone reception DB.

In one embodiment of the present invention, the subject of the plurality of categories of symptoms is selected from the group consisting of headache, chest pain, abdominal pain, back pain, labor pain, other pain, conscious disorder, airway water, dyspnea, respiratory arrest, , Seizures, syncope, nausea or vomiting, diarrhea or constipation, dysuria, hemoptysis, hematochezia, hemorrhage, vaginal bleeding, other bleeding, hyperthermia, hypothermia, dizziness or dizziness, paralysis, general weakness, mental disorders, Fracture, and the emergency dispatch includes a report consisting of neck, strangulation, suffocation, and drowning; And a patient state consisting of whether or not breathing and consciousness are made. The personal information includes the identity of the applicant, gender, age under 4 years, age 5 to 14, age 15 to 39, age 40 to 69 And age divided into 70 years or older.

In one embodiment of the present invention, the DB is included in each of the solutions, and provides individual information to the components of the respective solutions, and newly acquired information in the solution is stored in the DB.

[ Example ]

Hereinafter, an experiment according to an embodiment of the present invention will be described. Sixty - six cases were identified in which 6,600 cases were reported for the experiment and the main symptom and emergency patient classification were identified in connection with the emergency activity log and the National Emergency Department Information System (NEDIS). All of the 68 reported calls were revised. Since the personal information was deleted, the company correction work was performed by adding the virtual address to the address of the actual company and correcting it to include the patient information naturally. Patient information revision should be modified to include two pieces of information in the form of 'what template is in the template state'. This is because the semantic extraction is based on the Lebenstein distance between recognition results and templates.

The revised enterprise script consisted of 818 sentences in 68 scenarios, which were recorded and used as test input voice files. Experiments were carried out to test the recognition rate using 818 speech files as inputs and to test the performance of emergency dispatch by inputting 68 speech scenarios. Experiments were conducted with the scenario of FIG. 6 where an emergency dispatch was automatically detected. Figure 6 shows a flowchart of an emergency dispatch scenario, in accordance with one embodiment of the present invention. The output result is output in three kinds of emergency / non-emergency / prediction failure. The system was designed to output when the prediction failure failed to extract the patient information, or when there was no patient information for the emergency / non-emergency determination of the input data.

The complexity of the test input file is about 74 and the speech recognition result is as shown in [Table 5].

[Table 5] Recognition performance

Experimental results show that the word error rate is 9.14% and the sentence error rate is 24.21%. This is because the characteristics of the dialogue speech and the word segmentation method are used in Korean. When we look at the sentence recognition rate and the word error type below, it seems to be enough to extract the meaning from the recognition result.

Table 6 shows the detection performance of Emergency Emergency Situation in this experiment.

[Table 6] Emergency Emergency Detection Performance

Emergency dispatch detection performance is directly related to semantic extraction performance. This is because the recognition result and the emergency dispatch template are compared and analyzed to extract the meaning and the emergency dispatch is judged. When the sentence recognition rate is considered together with the sentence recognition rate (Table 5), the emergency dispatch detection performance is encouraging. In Table 6, the prediction failure is not judged by the system when the system lacks sufficient patient information or semantic information extraction. This is a case where there is insufficient patient information required by the system in the call contents because it is based on actual recording transcription.

The description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features presented herein.

Claims (7)

Emergency situation control system using Korean speech recognition technology with computer solution including display and DB,
The system comprises:
An automatic input unit for automatically inputting basic data by using Korean speech recognition, a semantic extraction unit for extracting meaning, and a multi-criteria severity classification solution, a multi-dispatch system command solution, and an emergency A voice recognition based report accepting solution having an automatic classification section for providing to a consulting solution;
A multi-criteria severity classification solution in which the severity of patients is classified based on the information received from the report acceptance solution and displayed on the display;
A multi-tasking system command solution for instructing the emergency dispatch based on the information received from the report acceptance solution and displaying on the display to provide necessary emergency action to the emergency personnel; And
And an emergency rescue consulting solution displayed on the display so as to guide the emergency care agent before the emergency agent on the basis of the information received from the report acceptance solution,
The information extracting unit of the speech recognition-based report acceptance solution compares the sentence of the situation room voice recognition result with the template of the core recognition object using the Lebenschin distance measurement model defined as Equation 1, , And personal information are extracted. The Lebenstein distance ratio is calculated using the Lebenstein distance measurement between the two strings a and b in Equation 1, and the threshold value of the key recognition target is obtained. Wherein the extraction of the key recognition target is determined by comparing and analyzing the speech recognition result with the speech recognition system.
[Formula 1]

(Formula 1)

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, and if it is different, the value is equal to 2)
[Formula 2]

(Where | a | and | b | are the lengths of the two strings) The method according to claim 1,
The main symptom of the patient is a headache, chest pain, abdominal pain, back pain, labor pain, other pain, conscious disorder, airway water, dyspnea, respiratory arrest, palpitations, cardiac arrest, convulsions, seizures, syncope, nausea or vomiting, diarrhea or constipation , Which is divided into 31 categories: urinary disorders, hemoptysis, hematochezia, nonhemorrhage, vaginal bleeding, other bleeding, hyperthermia, hypothermia, dizziness or dizziness, paralysis, general weakness, mental disorders,
Emergency control system using Korean speech recognition technology.
The method according to claim 1,
The emergency dispatch includes a report consisting of hanging, strangulation, suffocation and drowning; And
The patient's condition, which consists of whether or not breathing and consciousness,
Emergency control system using Korean speech recognition technology.
The method according to claim 1,
The personal information shall include the identity of the applicant, the sex, and the age divided into 4 and under, 5 to 14, 15 to 39, 40 to 69, and 70,
Emergency control system using Korean speech recognition technology.
The method according to claim 1,
Wherein the basic data automatic input unit of the voice recognition based report acceptance solution is a voice recognizer composed of an acoustic model, a language model, and a pronunciation model,
The acoustic model is a context-dependent triphone (triphone) model,
In the language model, words constituting sentences are divided into 1-gram (Unigram), 2-gram (Bigram) grouped by two consecutive words, 3-gram (Trigram) A 4-gram grouped by five, and a 5-gram grouped by five consecutive,
The pronunciation model is generated by a predetermined tone generator,
Emergency control system using Korean speech recognition technology.
As an emergency control method using Korean speech recognition technology,
The method comprises:
A step of automatically inputting basic data into a speech recognizer including an automatic input adding acoustic model, a language model, and a pronunciation model of a speech recognition based report acceptance solution provided in an emergency situation control system having a computer solution including a DB;
The information extracting section of the report acceptance solution compares the sentence of the situation room speech recognition result with the template of the core recognition object by using the Lebenschein distance measurement model in which the basic data is defined as follows: And Lebenstein distance ratios according to Equation 2 are obtained by using the Lebenstein distance measurement values between the two strings a and b of Equation 1 and the threshold value A semantic extraction step of determining a key recognition target extraction;
Providing the emergency classification information obtained from the extracted meaning to the multi-criteria severity classification solution, the multiple dispatch system command solution, and the emergency response consulting solution provided in the emergency management system;
Classifying the degree of severity of the patient based on the information provided by the multi-criteria severity classification solution from the report acceptance solution and displaying on the display;
Displaying on the display the multi-dispatch system command solution instructing the emergency dispatch based on the information provided from the report acceptance solution and providing the necessary emergency response to the first-aid agent; And
And displaying on the display the first aid consultation solution to guide the first aid agent before arrival of the emergency agent on the basis of the information provided from the report acceptance solution.
[Formula 1]

In Equation 1,

If the i-th character of the a-string is the same as the j-th character of the b-string

), The value is equal to 0, otherwise the value is equal to 2.
[Formula 2]

(Where | a | and | b | are the lengths of the two strings) The method according to claim 6,
The acoustic model is a context-based triphone (Triphone) model. In the language model, words constituting a sentence are divided into 1-gram (Unigram), 2-gram Bigram), a 3-gram (Trigram) group consisting of 3 consecutive 3-grams, a 4-gram group consisting of 4 consecutive 4 consecutive 5 consecutive 5 consecutive 5-gram consonants, Emergency situation control method using speech recognition technology.

Images