KR101923960B1 - Method for generating transcription from sound data based on voice recognition - Google Patents

Abstract

The present invention relates to a method for generating a transcription from sound data based on voice recognition and an apparatus using the same. In particular, according to the method of the present invention, a final transcription is generated from a primitive transcription by preforming a predetermined function corresponding to a functional word detected in the primitive transcription by a computing device. Accordingly, the present invention can promote convenience on the voice recognition.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for generating transcription statements from acoustic data based on speech recognition,

The present invention relates to a method for generating transcription from acoustic data based on speech recognition and an apparatus using the same. Specifically, according to the method of the present invention, a computing device performs a predetermined function corresponding to a functional word detected in a coarse transcription, thereby generating a final transcription from the coarse transcription.

BACKGROUND OF THE INVENTION [0002] Automatic speech recognition systems are becoming increasingly popular in the art of generating scripts, or transcription, from sound data. Although it is a general form to use an input device such as a keyboard in document creation, there may be a situation where the hand is inconvenient in the working environment, so that it is desired to use speech recognition as a means for overcoming this or as a means for increasing the input speed.

However, there are still many difficulties with such techniques, for example, when an automatic speech recognition system generates a primitive transcription in a transcription file, the transcriber modifies the primitive transcription to produce the final transcription. In this process, the transcriptionist hears a lot of voices, and it is easy to be tired to check carefully whether the punctuation marks are correctly printed on all sentences and whether the lines are properly spaced.

Therefore, according to the present invention, there is proposed a method of distinguishing functional words expressing a part of a transfer inquiry to be checked by a transcriber, in particular, a format of a document, which is not a voice to be characterized, thereby reducing the work of the transcriber, and a device using the method I want to.

US 2015-0142428 A1

An object of the present invention is to provide a method for efficiently creating a formatted document through voice input.

Specifically, the present invention aims to identify a function word that is distinguished from a voice to be text, thereby enabling to control the format and the like of a document.

As a result, the present invention aims at enhancing efficiency in transcription, thereby enabling transcriptionists to quickly and accurately produce high quality transcription doors.

The characteristic configuration of the present invention for achieving the object of the present invention as described above and realizing the characteristic effects of the present invention described below is as follows.

According to an aspect of the present invention there is provided a method of generating a transcription from acoustic data based on speech recognition, the method comprising the steps of: (a) step; (b) the computing device recognizing a voice from the acquired sound data to generate a primitive transcription statement; (c) detecting, by the computing device, a previously reserved function word with reference to a functional word dictionary in a generated primitive transfer statement; And (d) generating, by the computing device, a final transfer statement from the nested transfer statement by performing a predetermined function corresponding to the detected function word.

According to another aspect of the present invention, there is also provided a computer program stored in a machine readable non-transitory medium, comprising instructions embodied to perform the method according to the invention.

According to still another aspect of the present invention, there is provided a computing device for generating a transcription statement from acoustic data based on speech recognition, the apparatus comprising: a communication unit for acquiring the acoustic data; And (i) recognizing speech from the acquired sound data to generate a premise statement; (ii) a process of detecting a function word reserved in advance by referring to the functional word dictionary in the generated generic transcription statement; And (iii) performing a process to generate a final transfer statement from the nested transfer statement by performing a predetermined function corresponding to the detected function word.

According to the present invention, when the transcription sent by the trained artificial intelligence model is acquired, the part corresponding to the functional word can be identified, and the format of the document, such as punctuation, can be controlled, There is an effect.

In addition, according to the present invention, there is an effect that it is possible to assist transferors to produce high quality transfer data by increasing transfer efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, The figures can be obtained.
1 is a conceptual diagram schematically showing an exemplary configuration of a computing device that performs a method of generating a transcription statement from acoustic data based on speech recognition (hereinafter referred to as a "speech recognition transcription generation method") according to the present invention .
2 is an exemplary block diagram illustrating the hardware or software components of a computing device performing the method of generating a speech recognition transcription according to the present invention.
3 is a flowchart illustrating a method of generating a speech recognition transfer statement according to the present invention.

The following detailed description of the invention refers to the accompanying drawings, which illustrate, by way of example, specific embodiments in which the invention may be practiced in order to clarify the objects, technical solutions and advantages of the invention. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention.

The term " acoustic data " or " acoustic file " used throughout the description and claims of the present invention is a term that refers to a digital representation of acoustic information. For example, any acoustic or video format known to a person skilled in the art for acoustic data or sound files may be used. In some embodiments, the sound file may include a format that supports one audio codec, and in some embodiments, the sound file may include a format that supports multiple codecs. In some embodiments, the sound file may include an uncompressed audio format, such as WAV, AIFF, AU. In some embodiments, the audio file format is FLAC, Monkey's Audio with APE as file extension, WayPack with WV as file extension, Tom's loseless Audio Kompressor (TAK), TTA, ATRAC Advanced Lossless, Apple Lossless, and lossless WINDOWS media audio WMA). ≪ / RTI > In some embodiments, the format of the audio file may include lossy compression schemes such as MP3, Vorbis, Musepack, ATRAC, lossy WINDOWS media audio (WMA) and AAC.

In some embodiments, the acoustic format is a uncompressed PCM format such as " .wav " for WINDOWS computer readable media, and " .aiff " In some embodiments, Broadcast Wave Format (BWF) may be used, and the metadata may be stored together in a file. In some embodiments, the audio format is a lossless audio format such as FLAC, WayPack, Monkey's Audio, ALAC / Apple Lossless. In some embodiments, the lossless audio format provides a compression ratio of about 2: 1. In some embodiments, the audio format is a free-and-open format such as wav, ogg, mpc, flac, aiff, raw, au or mid. In some embodiments, the audio format is a published file format such as gsm, dct, vox, acc, mp4 / m4a, or mmf. In some embodiments, the audio format may be any suitable format such as mp3, wma, atrac, ra, ram, dss, msv, dvg, IVS, m4p, iklax, mxp4,

Throughout the detailed description and claims of the present invention, 'learning' or 'learning' refers to performing machine learning through computing according to a procedure, It is not intended to refer to training, which is used in a generally accepted sense of machine learning.

And throughout the description and claims of this invention, the word 'comprise' and variations thereof are not intended to exclude other technical features, additions, elements or steps. Also, 'one' or 'one' is used in more than one meaning, and 'another' is limited to at least the second.

Other objects, advantages and features of the present invention will become apparent to those skilled in the art from this description, and in part from the practice of the invention. The following examples and figures are provided by way of illustration and are not intended to limit the invention. Accordingly, the details disclosed herein with respect to a particular structure or function are not to be construed in a limiting sense, but merely as being representative of the general inventive concept providing a guideline for carrying out the invention in various detail structures, It should be interpreted as basic data.

Moreover, the present invention encompasses all possible combinations of embodiments shown herein. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, certain features, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in connection with one embodiment. It should also be understood that the position or arrangement of individual components within each disclosed embodiment may be varied without departing from the spirit and scope of the present invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is to be limited only by the appended claims, along with the full scope of equivalents to which such claims are entitled, if properly explained. In the drawings, like reference numerals refer to the same or similar functions throughout the several views.

Unless otherwise indicated herein or clearly contradicted by context, items referred to in the singular are intended to encompass a plurality unless otherwise specified in the context. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, so that those skilled in the art can easily carry out the present invention.

1 is a conceptual diagram schematically illustrating an exemplary configuration of a computing device that performs a method for generating a speech recognition transcription of the present invention.

1, a computing device 100 according to an embodiment of the present invention includes a communication unit 110 and a processor 120. The communication unit 110 communicates with an external computing device (not shown) Communication is possible.

In particular, the computing device 100 may be implemented as a computer-readable medium, such as conventional computer hardware (e.g., a computer processor, memory, storage, input and output devices, Electronic communication devices, electronic information storage systems such as network-attached storage (NAS) and storage area networks (SAN), and computer software (i.e., computing devices that enable a computing device to function in a particular manner) Commands) to achieve the desired system performance.

The communication unit 110 of the computing device can send and receive requests and responses to and from other interworking computing devices. As an example, such requests and responses can be made by the same transmission control protocol (TCP) session But not limited to, a user datagram protocol (UDP) datagram, for example. In addition, in a broad sense, the communication unit 110 may include a keyboard, a mouse, an external input device, a printer, a display, and other external output devices for receiving commands or instructions.

The processor 120 of the computing device may also be a micro processing unit (MPU), a central processing unit (CPU), a graphics processing unit (GPU), a neural processing unit (NPU) or a tensor processing unit ), A data bus, and the like. It may further include a software configuration of an operating system and an application that performs a specific purpose.

2 is an exemplary block diagram illustrating the hardware or software components of a computing device performing the method of generating a speech recognition transcription according to the present invention.

Referring briefly to the configuration of a method and apparatus in accordance with the present invention with reference to FIG. 2, the computing device 100 may include an acoustic data input 210 as a component thereof. The sound data input unit 210 acquires sound data according to the method of the present invention. The voice recognition unit 220 receiving the acquired sound data is configured to recognize the voice from the sound data and generate the premise message. The voice recognition unit 220 includes a sound data input unit 210 and a voice recognition unit 220 2 may be implemented by interworking of the communication unit 110 or the processor 120 included in the computing device 100 or the communication unit 110 and the processor 120, The technician will understand.

The primitive transcription is the result of transcription from the audio data, which can be passed to the functional word detector 230 configured to detect the functional word. The computing device 100 may include a functional word dictionary unit 250 as a component for providing a functional word for detection to the functional word detection unit 230. The functional word dictionary unit 250 may store the functional word itself, (Not shown) interlocked with each other via the network 110 may have a functional word. In the latter case, the other computing device can manage the functional word according to a request and a response of the functional word dictionary unit 250.

Next, the functional word processing unit 240 is configured to perform a predetermined function corresponding to the functional word detected by the functional word detection unit 230 to generate a final transfer statement from the rough transfer statement. The external entity may be a medical staff member including a user of the computing device 100, for example, a doctor, a nurse, or the like, a manager other than the manager, But it should be understood that any subject that includes the detection of errors in the result of the transcription should be understood as being included.

The specific functions and effects achieved by the components shown in FIG. 2 will now be described in detail with reference to FIG. Although the components shown in FIG. 2 are illustrated as being realized in one computing device for convenience of explanation, it will be understood that the computing device 100 performing the method of the present invention may be configured such that a plurality of devices are interlocked with each other. E.g,

The sound data input unit 210, the speech recognition unit 220, the functional word detection unit 230, the functional word processing unit 240 and the functional word editing unit 260 are implemented in the first computing device and have functional words necessary for detecting and processing the functional word, The functional word dictionary unit 250 to be managed may be configured to be executed in a client connected to the server. It is possible for a typical technician to configure the plurality of devices to interwork with each other in various ways to perform the method of the present invention will be.

An embodiment of a method for generating a speech recognition transfer statement according to the present invention will now be described in more detail with reference to FIG. 3 is a flowchart illustrating a method of generating a speech recognition transfer statement according to the present invention.

Referring to FIG. 3, the method for generating a speech recognition transfer statement according to the present invention includes a step S 100 of acquiring sound data by an acoustic data input unit 210 implemented by the computing device 100.

Next, a method for generating a speech recognition transfer statement according to the present invention comprises the steps of: a voice recognition unit 220 implemented by the computing device 100 recognizes a voice from the obtained sound data to generate a primitive transfer statement S200). Generation of transcripts can be done by a machine learning model, for example, in-depth neural networks can be used. It is known to those of ordinary skill in the art that the in-depth neural network may consist of an acoustic model and a language model. The detailed description of the acoustic model and / or the construction of the language model will be omitted in order not to obscure the details of the present invention.

When the generic transcription is generated in step S200, the functional word detection unit 230 implemented by the computing device 100 generates a generic transcription statement by referring to the functional word dictionary 250 in the generated generic transcription statement, (S300). ≪ / RTI >

Here, the term "functional word" is a command distinguished from a character string to be characterized, and is a vocabulary that enables to efficiently formulate a formatted document by mainly describing and defining the format of the document. Preferably, the functional word may be constructed of a commonly used natural language.

Today's weather is clear.
What about the weather tomorrow?

For example, in order to input the contents of Table 1 above, if you input the voice saying, "Today's weather is clear, change the line and change the line and tomorrow's weather will be a question mark," you can say "Today's weather is clear" and " The function is a function that means to perform "function to take a period", "function to continue writing on the next line", "function to take a question mark", and "function to take a period", " There are functional words corresponding to simple substitution, function words corresponding to syllabus expansion, and functional words corresponding to the use of representative words. Examples of functional words include simple substitution, slang extension, .

Among these, simple substitution refers to a function of replacing a corresponding function word with a specific string. For example, "period", "period", "period" is a function word for representing a period (.), Which can be replaced by a period (.).

For example, an example objc code for handling a simple substitution when a function word of a period and a line-feed function following the period is provided may be as follows. This is to handle two consecutive simple substitution functional words, 'Replace Period'.

if ([self.words [i] containsPeriodCommand] && [self.words [i + 1] containsLineFeedCommand]) {
[outputString appendString: @ ". \ n"];
i + = 2;
}

In addition, the slave language extension refers to the function of deleting the corresponding functional word in relation to other functional words depending on the position. This is an extension type that includes a predicate added in a language habit without any meaning. It is a form of extension that includes "punctuation", "let", "period", "period" , "Do", "insert", "take", "do", "do" and "insert" when they are present are excluded from the actual string, . This example can be shown in Table 3.

[Period, period, period] {take, do, put}
Here, one of the periods, periods, or periods in [] is required, and predicates in {} are classified as functional words only if they exist.

As shown in Table 3, the syllabary expansion may include each expected item, and it is also possible to separately manage it as shown in Table 4 below and replace it with the representative language.

[Period, period, period] {$ swordsman}
Sailor: Take, do, do, put, put

Next, the use of a representative word refers to a function of replacing a character string of a predetermined format with a character string of another format. For example, you can use a delegate to represent a particular type of string, such as a number. If you enter a number such as "small 1", "parentheses 2", and so on, . In this case, it can be expressed as "[small, parentheses] $ int times", "small one" can be replaced with "1)", and "parentheses 2" can be replaced with "2)".

As another example, it is also possible to replace "$ int1 - $ int2 - month $ int3 - day" with "$ int1 - $ int2 - $ int3". For example, "December 25, 2018" may be replaced with "2018-12-25" by the use of representative words.

As another example, the functional group of functional words may further include template substitution. This template substitution refers to the function of loading a predetermined template and inserting an individual string into at least one specified position of the template. The loading of the template means a function of loading a template of a medical certificate in a predetermined form through a functional word such as " format is a certificate of diagnosis ", and a function of inserting an individual character string is a function word such as & Means a function of inputting information corresponding to an item which is essential or optional in a medical certificate. In order to complete the template, a functional word such as " Completion of medical certificate creation " may be additionally requested.

The functional word is not limited to the example described above, and may include, for example, excluding the functional word. This function word exclusion refers to a function for preventing a function corresponding to a function word inputted during the application from being performed so that the inputted function word is included in the final transfer statement.

For example, if a functional word with the function word exclusion function is set to " Exclude function word", a voice "One function words exclude one word" can be issued to input a sentence of "Change one word". Also, if you want to enter "Exclude one function word", you can say "Exclude function word exclusion function".

If it is necessary to continue to exclude such function words continuously, it may be configured to turn on / off the application of the function word exclusion to the "start of excluding function words" and "end of function words" (on / off toggle).

3, a method for generating a speech recognition transfer statement according to the present invention is a method for generating a speech recognition transfer statement according to an embodiment of the present invention. In the method for generating a speech recognition transfer statement according to the present invention, a function word processing unit 240 implemented by the computing device 100 performs a predetermined function corresponding to the detected function word, (S400) of generating a final transfer statement from the statement, and the predetermined function is exemplarily described above.

The functional word is held in the functional word dictionary unit 250 and may be added, modified or deleted by the user as needed. To this end, the method for generating a speech recognition transfer statement according to the present invention is a method for generating a speech recognition transfer statement according to an embodiment of the present invention, in which a functional word editing unit 260 implemented by the computing device 100 performs a function corresponding to a specific functional word And performing at least one of editing, deleting, and adding (S500). In FIG. 3, each step of the method according to the present invention is shown to be performed in the order of S100, S200, S300, S400, which is a logically inevitable order, while the order of S500 is concurrent with each step, or It will be understood by those skilled in the art that this may be done at this time and that S500 is not meant to be performed after S200.

Specifically, in operation S500, in response to an editing request for the functional word, the functional word editing unit 260 implemented by the computing device reads (i) a specific functional word included in the functional word dictionary, (S520) of providing a predetermined user interface configured to enable at least one of editing, deleting, and (ii) adding a specific function word not included in the functional word dictionary and a function to the specific function word; And a step (S540) of performing at least one of the editing, deleting, and adding the function word editing unit 260 implemented by the computing device in response to the operation on the user interface.

By carrying out the above-described steps of the present invention, it is possible to create transcription statements having a more appropriate format by utilizing functional words based on speech recognition. Steps S100 to S400 may be performed only once, but it is preferable that real-time processing is performed on at least a part of the continuously inputted sound data so as to correspond to sequential speech recognition of sequentially inputted sound data, time and / or iteratively.

As described so far, the present invention is able to utilize functional words other than the transferred character strings throughout all the embodiments and modifications, thereby making it possible to complete a transfer style of a desired proper format.

Based on the description of the embodiments above, those skilled in the art will recognize that the methods and / or processes of the present invention and their steps may be implemented in hardware, software, or any combination of hardware and software suitable for the particular application Points can be clearly understood. The hardware may include special features or components of a general purpose computer and / or a dedicated computing device or a specific computing device or a particular computing device. The processes may be realized by one or more microprocessors, microcontrollers, embedded microcontrollers, programmable digital signal processors or other programmable devices having internal and / or external memory. Additionally or alternatively, the processes can be configured to process application specific integrated circuits (ASICs), programmable gate arrays, programmable array logic (PAL) Or any other device or combination of devices. Furthermore, the objects of the technical solution of the present invention, or portions contributing to the prior art, may be implemented in the form of program instructions that can be executed through various computer components and recorded on a machine-readable recording medium. The machine-readable recording medium may include program commands, data files, data structures, and the like, alone or in combination. The program instructions recorded on the machine-readable recording medium may be those specially designed and constructed for the present invention or may be those known to those of ordinary skill in the computer software arts. Examples of the machine-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tape, optical recording media such as CD-ROM, DVD, Blu-ray, magneto-optical media such as floptical disks magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include, but are not limited to, any of the above devices, as well as a heterogeneous combination of processors, processor architectures or combinations of different hardware and software, Which may be constructed using a structured programming language such as C, an object-oriented programming language such as C ++ or an advanced or low-level programming language (assembly language, hardware description languages and database programming languages and techniques) This includes not only bytecode, but also high-level language code that can be executed by a computer using an interpreter or the like.

Thus, in one aspect of the present invention, when the methods and combinations described above are performed by one or more computing devices, combinations of the methods and methods may be implemented as executable code that performs each of the steps. In another aspect, the method may be implemented as systems for performing the steps, and the methods may be distributed in various ways throughout the devices, or all functions may be integrated into one dedicated, stand-alone device, or other hardware. In yet another aspect, the means for performing the steps associated with the processes described above may include any of the hardware and / or software described above. All such sequential combinations and combinations are intended to be within the scope of this disclosure.

For example, the hardware device may be configured to operate as one or more software modules to perform processing in accordance with the present invention, and vice versa. The hardware device may include a processor, such as an MPU, CPU, GPU, TPU, coupled to a memory, such as ROM / RAM, for storing program instructions and configured to execute instructions stored in the memory, And a communication unit capable of receiving and sending data. In addition, the hardware device may include a keyboard, a mouse, and other external input devices for receiving commands generated by the developers.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, Those skilled in the art will appreciate that various modifications and changes may be made thereto without departing from the scope of the present invention.

Therefore, the spirit of the present invention should not be construed as being limited to the above-described embodiments, and all of the equivalents or equivalents of the claims, as well as the following claims, I will say.

Such equally or equivalently modified means include, for example, a logically equivalent method which can produce the same result as the method according to the present invention, Should not be limited by the foregoing examples, but should be understood in the broadest sense permissible by law.

Claims (12)

A method for generating a transcription from acoustic data based on speech recognition,
(a) the computing device acquiring the sound data;
(b) the computing device recognizing a voice from the acquired sound data to generate a primitive transcription statement;
(c) detecting, by the computing device, a previously reserved function word with reference to a functional word dictionary in a generated primitive transfer statement; And
(d) generating, by the computing device, a final transfer statement from the primitive transfer statement by performing a predetermined function corresponding to the detected function word
Lt; / RTI >
Wherein the functional word corresponds to at least one function belonging to a function group including simple substitution and syllabary expansion,
In the simple substitution,
And replacing the corresponding function word with a specific character string,
The syllabic expansion may include:
And deleting the corresponding function word in relation to other function words depending on the position. The method according to claim 1,
(e) performing at least one of editing, deletion and addition of a specific function word in the functional word dictionary and a function corresponding to the specific functional word according to an editing request for the functional word
Further comprising the steps of: 3. The method of claim 2,
The step (e)
(e1) in response to an editing request for the functional word, the computing device performs: (i) editing and deleting a specific functional word included in the functional word dictionary and a function for the specific functional word; and (ii) Providing a predetermined user interface configured to enable at least one of a specific function word included and an addition of a function to the specific function word; And
(e2) in response to an operation on the user interface, the computing device performing at least one of the editing, deleting, and adding
And generating a speech recognition transcription statement. delete The method according to claim 1,
The functional group further includes a representative word utilization,
The above-
And replacing the character string of the predetermined format with a character string of another format. The method according to claim 1,
Wherein the functional group further comprises template substitution,
The template substitution may include:
And a function of loading a predetermined template and inserting an individual character string into at least one predetermined position of the template. The method according to claim 1,
Wherein the functional group further includes functional word exclusion,
In the case of excluding the functional word,
Wherein a function corresponding to the input functional word is not performed so that the functional word input during the exclusion of the functional word is included in the final transcriptional statement. The method according to claim 1,
Wherein the sound data is inputted continuously and the steps (a) to (d) are performed real-time and iteratively for at least a part of the sequentially inputted sound data A method for generating a speech recognition transcription. A computer program product, stored in a machine readable non-volatile storage medium, comprising instructions embodied in a computer-readable medium for causing a computing device to perform the method of any one of claims 1 to 3, . An apparatus for generating transcription from acoustic data based on speech recognition, the apparatus comprising:
A communication unit for acquiring the sound data; And
(i) recognizing speech from the acquired sound data and generating a premise word; (ii) a process of detecting a function word reserved in advance by referring to the functional word dictionary in the generated generic transcription statement; And (iii) a processor for performing a process of generating a final transfer statement from the nested transfer statement by performing a predetermined function corresponding to the detected function word
Lt; / RTI >
Wherein the functional word corresponds to at least one function belonging to a function group including simple substitution and syllabary expansion,
In the simple substitution,
And replacing the corresponding function word with a specific character string,
The syllabic expansion may include:
And deleting the corresponding function word in relation to other function words depending on the position. 11. The method of claim 10,
The processor comprising:
Deletes and adds a function corresponding to the specific function word and the specific function word in the function word dictionary according to an edit request for the function word. 12. The method of claim 11,
The processor comprising:
(iv) editing a specific function word included in the functional word dictionary and a function for the specific functional word, and deleting the specific functional word included in the functional word dictionary and the function for the specific functional word in response to an editing request for the functional word The method comprising the steps of: And (v) performing at least one of the editing, deleting, and adding in response to an operation on the user interface.

Images