Foreign language speaking learning system using image-associative characters and artificial intelligence and foreign language speaking learning method using the same

Abstract

The present invention relates to a foreign language speaking learning system utilizing image association characters and artificial intelligence, and a foreign language speaking learning method using the same. More specifically, it relates to a foreign language speaking learning system utilizing image association characters and artificial intelligence and a foreign language speaking learning method using the same, which improves memorization ability regarding words and sentence structures through an image association memory method by remembering words and sentence structures through image association characters that convert words into specific images or abbreviations when remembering words or sentence structures during the process of learning a foreign language other than one's native language, and can improve speaking and speaking ability by detecting words spoken by the learner and determining whether they are correct.

Description

Technology Field

[0001] The present invention relates to a foreign language speaking learning system utilizing image association characters and artificial intelligence, and a foreign language speaking learning method using the same. More specifically, it relates to a foreign language speaking learning system utilizing image association characters and artificial intelligence and a foreign language speaking learning method using the same, which improves memorization ability regarding words and sentence structures through an image association memory method by remembering words and sentence structures through image association characters that convert words into specific images or abbreviations when remembering words or sentence structures during the process of learning a foreign language other than one's native language, and can improve speaking and speaking ability by detecting words spoken by the learner and determining whether they are correct. Background Technology

[0002] In terms of foreign language learning methods, traditional approaches utilize five-sense skills such as sound, writing, and repetition, which involve memorizing by continuously repeating a single word or sentence while reading it out loud or writing it down.

[0003] However, while these traditional learning methods are effective for short-term memory, they have the problem of being difficult to transfer learning results to long-term memory, and fixed learning methods have the drawback of easily exhausting the brain.

[0004] Here, short-term memory and long-term memory relate to human memory; long-term memory includes episodic memory, which involves recalling memories; semantic memory, which involves retaining knowledge; and procedural memory, which is the memory of the sequence of actions memorized through the body.

[0005] In long-term memory, episodic and procedural memories are more intense and last longer than semantic memories. This is because episodic memories are generally accompanied by emotions, which provide stronger stimulation to the brain, while procedural memories utilize a wider area of ​​the brain.

[0006] Therefore, when learning a foreign language, it is more desirable to use a method that leads to long-term memory rather than relying on short-term memory through traditional repetitive learning.

[0007] In particular, learning a foreign language is often likened to memorizing words, and various word memorization methods have been proposed for this purpose.

[0008] For example, the image association memory method, also known as the Roman Room memory method, is known to have been handed down from the Roman era and is a method of creating a memory room in one's mind using familiar places and objects within them, and then associating the content one wants to memorize with each of the objects.

[0009] However, the image association method, which involves matching sentences or content to be remembered with related objects, has the problem of not being suitable for memorizing words. Furthermore, when associating one object with each word, the number of objects increases as the number of words to be remembered increases; consequently, not only are there few objects available for actual memory, but there are also many cases where only the word represented by a single object is remembered, making it unsuitable for learning words or sentences.

[0010] In addition, Korean Published Patent No. 10-2014-0077144 proposed an 'English education method and online game method using chunks,' and Korean Published Patent No. 10-2010-0078375 proposed a 'word matching game method using speech recognition technology.'

[0011] Among the aforementioned prior art, Korean Published Patent No. 10-2010-0078375 includes a step of displaying a selected set of words with native speaker pronunciation and having the user memorize them for a certain period of time, and a speech recognition step of collecting user pronunciation data, performing speech recognition, and notifying the speech recognition result.

[0012] The aforementioned conventional technology involves providing a word selected according to the user's request and simultaneously pronouncing the word with native pronunciation to enable the user to learn, thereby allowing the user to memorize words by utilizing both eyes and ears. However, this method of memorization is merely an addition of native pronunciation to the aforementioned traditional learning method, and thus, it could not be expected to have a significant effect in improving word memorization ability. Prior art literature

[0013] Republic of Korea Published Patent No. 10-2014-0077144 Republic of Korea Published Patent No. 10-2010-0078375 The problem to be solved

[0014] The present invention aims to provide a foreign language speaking learning system utilizing image association characters and artificial intelligence, which solves the above-mentioned problems and, in particular, enhances memorization effectiveness by enabling learners to connect words and sentences to long-term memory during the process of memorizing and learning, as well as a foreign language speaking learning method using the same.

[0016] In particular, the present invention aims to provide a foreign language speaking learning system utilizing image association characters and artificial intelligence, and a foreign language speaking learning method using the same, wherein the learner is provided with image association characters consisting of specific images or abbreviations for each word, and together with this, a native speaker pronounces and provides a learning text, thereby recognizing the image association characters composed of images or abbreviations as well as visual and auditory senses, and associative memory is achieved by substituting the image association characters into the original text to remember them, so that the word and sentence structure can be led into long-term memory.

[0018] In addition, the present invention aims to provide a foreign language speaking learning system utilizing image association characters and artificial intelligence, and a foreign language speaking learning method using the same, which enables objective verification of a learner's speaking ability by receiving the sound of an answer spoken as a result of learning after the learner has completed learning a learning text, determining whether it is correct, and providing only the word determined to be correct as a word of the original text. means of solving the problem

[0019] To achieve the above objective, the present invention,

[0020] A database unit storing image association characters corresponding to one or more of specific images or abbreviations for each learning word;

[0021] A CAT conversion unit that, when an original text composed of one or more original learning words to be provided to a learner is input, extracts the original learning words included in the input original text, calls image association characters corresponding to the extracted words from the database unit and converts them into image association characters, generates a learning text composed of image association characters, classifies the generated learning text through one of the classification methods of difficulty, situation, or type, and then transmits it to the database unit for storage;

[0022] A passage provisioning unit that provides learning passages generated through the above-mentioned CAT conversion unit to a learner online or offline;

[0023] A learning provision unit that, after confirming the receipt of the learning text provided through the above-mentioned text providing unit, plays a recording file in which the above-mentioned original text is recorded with the original pronunciation previously stored by a native speaker or an instructor capable of providing a corresponding intonation and pronunciation;

[0024] It is characterized by including a learning confirmation unit that receives the sound of the learner speaking about the learning text after the learner's learning through the learning provision unit is completed, and compares it with the original text stored in advance to determine whether it is the correct answer.

[0026] In addition, the learning verification unit preferably includes a voice recognition unit that recognizes sounds uttered by a learner, a word conversion unit that converts the recognized sounds into words, a speaking verification unit that converts words determined to be correct answers into original learning words included in the original text by comparing the converted words with the correct answer text and provides them to the learner, and a repetitive listening unit that plays back sentences containing sounds that are not converted into words or words determined to be incorrect answers through the recording file.

[0028] In addition, when the above CAT conversion unit converts words included in the input original text into corresponding image association characters, it is desirable to convert one word including one or more image association characters, and the converted image association characters are provided to the learner according to a preset difficulty level, provided in order from lowest to highest difficulty level or according to the difficulty level set by the user.

[0030] Meanwhile, in a foreign language speaking learning method utilizing image association characters and artificial intelligence according to the present invention,

[0031] A step of matching original learning words with corresponding image association characters and storing them in a database;

[0032] Step of inputting text to be provided to the learner;

[0033] A step of generating a learning audio file by having an instructor record the input text with a native speaker or equivalent intonation, pronunciation, and speed;

[0034] A step of generating a learning passage converted into image association characters that match the learning words constituting the input passage, and saving the input passage as the correct answer passage;

[0035] A step of providing the generated learning text to the learner online via a telecommunications line or in a printed state on paper via an output device;

[0036] A step of playing the above-mentioned learning recording file after the learning text has been provided to the learner;

[0037] It is characterized by including a step of checking the answer written by the learner, determining whether the image association character is correct, and scoring.

[0039] In addition, the scoring step preferably includes a speech recognition step for recognizing sounds uttered by the learner, a word conversion step for converting the recognized sounds into words, a speaking verification step for providing the learner with words determined to be correct by comparing the converted words with the answer key, and a repeated listening step for playing back sentences containing sounds that are not converted into words or words determined to be incorrect through the recording file. Effects of the invention

[0040] The present invention, as described above, has the effect of improving memorization ability by leading to long-term memory through visual and auditory stimulation of the learner, by learning original words during the process of learning a foreign language using image association characters consisting of specific images or abbreviations and audio files recorded by a native speaker or an equivalent instructor.

[0042] In addition, the present invention can improve a learner's speaking and speaking ability by recognizing the answer voice written by a learner who has studied the learning text and the recording file, or the sound uttered by the learner, and determining whether it is correct.

[0043] At this time, the present invention has the advantage of improving learning efficiency by converting only the words judged to be correct among the sounds spoken by the learner into original words and providing them to the learner, and by repeatedly playing the recording file until the learner guesses all the original words constituting the learning text, thereby enabling repetitive learning. Brief explanation of the drawing

[0044] FIG. 1 is a schematic diagram illustrating a foreign language speaking learning system utilizing image association characters and artificial intelligence according to the present invention. FIG. 2 is a configuration diagram illustrating a foreign language speaking learning system utilizing image association characters and artificial intelligence according to the present invention. FIG. 3 is an exemplary diagram illustrating an example of a learning text and answer voice according to the present invention. FIG. 4 is an exemplary diagram illustrating another example of a learning text and answer voice according to the present invention. FIG. 5 is a configuration diagram illustrating the configuration of a CAT converter constituting the present invention. FIG. 6 is an example diagram illustrating a foreign language speaking learning method using image association characters and artificial intelligence according to the present invention. FIG. 7 is an example illustrating a novel word matching step constituting the present invention. FIG. 8 is a flowchart illustrating the operation flow of the new word matching step shown in FIG. 7. FIG. 9 is an example diagram illustrating a speaking judgment step constituting the present invention. Specific details for implementing the invention

[0045] Hereinafter, a foreign language speaking learning system utilizing image association characters and artificial intelligence according to the present invention and a foreign language speaking learning method using the same will be examined in detail with reference to the attached drawings.

[0047] First, the foreign language writing learning system (1) utilizing image association characters and artificial intelligence according to the present invention comprises, as illustrated in FIG. 1, a database unit (10) in which image association characters (CAT) corresponding to either a specific image or an abbreviation for each learning word are stored; a CAT conversion unit (20) which, when an original text to be provided to a learner is input, extracts words included in the input original text (OT), calls the image association characters (CAT) corresponding to the extracted words from the database unit (10) and converts them into image association characters, generates a learning text (CT) composed of image association characters, classifies the generated learning text (CT) through one of the classification methods of difficulty, situation, or type, and then transmits and stores it in the database unit (10); a text unit (30) that provides the learning text (CT) generated through the CAT conversion unit (20) to the learner online or offline; and after the receipt of the learning text (CT) provided through the text unit (30) is confirmed, the original text is provided to a native speaker or It includes a learning provision unit (40) that plays a recording file recorded with original pronunciation that has been saved in advance by an instructor capable of providing corresponding intonation and pronunciation, and a learning confirmation unit (50) that receives an answer voice (AT) generated in real-time or as a recording file of the sound the learner uttered regarding the learning text (CT) after the learner's learning through the learning provision unit (40) is completed, and compares it with the original text that has been saved to determine whether it is correct.

[0049] The above database unit (10) electronically stores image association characters (CAT) and stores them systematically, and it is preferable that it be managed by a database management system (DBMS), which is a system that manages data storage.

[0050] The database unit (10) of the present invention is composed of hardware and software that manages it, as described above. Since the specific structure and operational relationship of the database unit (10) are well known, a detailed description will be omitted.

[0051] In addition, the above database unit (10) is connected to the learner via a wired or wireless communication means through a cloud server as shown in FIG. 1, and enables the transmission and reception of learning texts, answer voices, and the learner's voice answers.

[0052] However, in the present invention, the database unit (10) stores image association characters (CAT) corresponding to words of the language that the learner intends to learn. In the following description, the language that the learner intends to learn is English, as described and illustrated as an example, but this is merely one example, and it is obvious that it can be utilized with any foreign language other than the learner's native language, such as Spanish, French, Chinese, Japanese, or Korean.

[0053] In the following description, learning English is described as an example.

[0054] The image association character (CAT) corresponding to the English word stored in the database unit (10) consists of either an image or an abbreviation. As is known, the number of words constituting English is approximately 600,000 as of 1989, based on the Oxford English Dictionary, but the number of words actually used by native speakers is known to be about 20,000. Teachers and instructors currently teaching English to students claim that they can understand conversations, newspaper readings, movies, etc., within the range of 90-95% using only about 3,000 English words.

[0055] Accordingly, the image association characters (CAT) stored in the database unit (10) in the present invention are formed by selecting 3,000 essential words from English words and converting each word into one or more of images or abbreviations.

[0056] In Figure 3, the text placed on the left is a learning text (CT) converted into image association characters (CAT), and the composition written on the right is an answer voice (AT) in which the learner wrote the text (CT) in English after learning it.

[0057] In Figure 3, the sentence written by the learner has a 100% correct answer rate for the learning text (CT). Based on this, an example of the image association character (CAT) according to the present invention is as follows.

[0058] First, image association characters ' refers to the protagonist 'Kelly's' in the original text.

[0059] A sentence composed of a combination of the following image association characters: ' is 'family goes to a mexican restaurant' in the original text.

[0060] In other words, image association characters ' ' is 'family', and is an image association character ' is 'goes', and the image association character '→' is 'to'. And the image association character ' ' is 'mexican', and is an image association character ' ' is 'restaurant' in the original text.

[0061] As described above, in the present invention, image association characters (CAT) are formed as images or abbreviations for each word constituting a sentence, and then these are constructed as a sentence identical to the sentence of the original text to generate a learning text (CT).

[0062] There are also words here that are not transformed into Image Association Characters (CAT), such as 'It's,' the first word of the very next sentence. 'It's' is a shortened form of the possessive pronoun 'It is,' and as such, it is a basic word of the lowest difficulty level, so it was not converted separately.

[0063] Meanwhile, looking at the following sentence, abbreviations and images are mixed, and the image association characters are ' And, the words in the original text are 'It's very popular and it's always full of people'.

[0064] Although some of the image association characters (CAT) of the present invention have been illustrated through the above description and Figure 3, these image association characters (CAT) are merely one embodiment and can be modified to improve the learning efficiency of learners.

[0065] For example, when providing a learning text (CT) composed of sentences converted into the image association characters (CAT) of the present invention to a learner, if the learner is an online learner connected via a telecommunications line, when the native speaker's pronunciation provided along with the learning text (CT) is emitted through a speaker, effects such as changing the color of the image association characters being read by the native speaker or enlarging and then returning the size of the image association characters may be provided. In other words, various additional effects may be provided depending on the learner's learning environment to enhance the learner's learning efficiency.

[0067] Next, the above CAT conversion unit (20), when an original text to be provided to a learner is input, extracts words included in the input original text, calls an image association character (CAT) corresponding to the extracted words from the database unit (10) and converts it into an image association character, generates a learning text (CT) composed of image association characters, classifies the generated learning text (CT) through one of the classification methods of difficulty, situation, and type, and then transmits it to the database unit (10) for storage.

[0069] The above CAT conversion unit (20) inputs an original text to be provided by a teacher or instructor preparing a lecture for the learner's learning, and when the original text is input, it analyzes the sentences included in the text and extracts the words that make up each sentence.

[0070] An image association character (CAT) corresponding to the extracted word is called from the database unit (10), and the called image association character (CAT) is arranged to have the same sentence structure as the original text, thereby generating a learning text (CT).

[0071] The generated learning text (CT) is provided to the learner online or offline, and is simultaneously transmitted to and stored in the database unit (10).

[0072] At this time, multiple image association characters (CATs) may correspond to one word constituting the original text, and preferably, when multiple image association characters (CATs) correspond to one original word, the image association characters (CATs) are composed of abbreviations.

[0073] In other words, abbreviations are shortened words that have a relatively smaller number of letters compared to the original word, and the probability of a learner getting the correct answer varies depending on the number of letters omitted from a single word.

[0074] Therefore, as an example of setting the difficulty level, the lowest difficulty level is one where the number of characters omitted is within the range of 1 to 3, the middle difficulty level is one where the number of characters omitted is 3 to 5, and the highest difficulty level is one where the number of characters omitted is 3 to 5, but the omitted characters are in non-consecutive positions.

[0075] As described above, when multiple image association characters (CATs) with different difficulty levels are formed for a single original word, the image association characters (CATs) that match the difficulty level set by the teacher or instructor are initially exposed to the learner during the process of converting the original text containing the original word into a learning text (CT).

[0076] Subsequently, when the difficulty level is changed at the request of a learner or by the teacher or instructor, the difficulty-adjustable Image Association Character (CAT) provides the learner with an Image Association Character (CAT) corresponding to the adjusted difficulty level to maintain the learner's interest in learning.

[0077] Meanwhile, when forming image association characters (CAT), if the input original word consists of an idiom (English idiom) which is a combination of two or more words used differently from the original meanings, a separate image association character (CAT) to express the idiom may be applied to the learning text (CT).

[0078] The present invention separately stores 100 idioms that are typically used in the CAT conversion unit (20), and when a word extracted from the original text is determined to be an idiom, calls the image association character (CAT) corresponding to the idiom and applies it to the learning text (CT).

[0079] In other words, as previously mentioned, idioms are combinations of two or more words and can be classified into preposition + noun, noun + preposition, idioms used in the form of verb + object + preposition using specific verbs and prepositions, idioms used in the form of verbs, and be verb + past passive voice; depending on the classification, representative multiple examples of these idioms are provided as separate image association characters.

[0080] At this time, the teacher or instructor can choose to provide the word recognized as an idiom as an idiom-specific image association character or as a general image association character.

[0081] In addition, for collocations that are fixed combinations of words in idioms and have the same meaning as combinations of individual words, one can choose to provide each word as an image association character or provide an image association character dedicated to the idiom.

[0082] As an example of the above idiom, 'free of charge' means 'free of charge', and 'free of charge' can be provided as image association characters designated as individual words 'fre', ' / ', and 'chag', or as another example, '0$' can be provided as an idiom-specific image association character created in image form.

[0083] Meanwhile, the CAT conversion unit (20) classifies the learning text (CT) according to one of the categories of difficulty, situation, and type when a teacher or instructor inputs the original text or after the learning text (CT) is generated, and then transmits it to the database unit (10) so that it is stored in the database unit (10) according to the classification.

[0084] The classification-based storage of these learning passages (CTs) enhances learners' interest by providing passages according to their desired difficulty, situation, and type, and improves learning efficiency by providing continuity of learning.

[0085] For example, as shown in FIG. 4, if the learning text (CT) stored according to classification by type is related to food, the learning text (CT) shown in FIG. 4 is provided as a recommended learning text (CT) to the learner who has studied the learning text (CT), thereby increasing learning efficiency by continuously studying food or learning texts related thereto.

[0086] Here, recommended learning passages (CT) are selected from among the learning passages stored by category in the aforementioned database unit (10) that are classified as the same type as the learning passage (CT) that the learner studied immediately before, and among the selected multiple learning passages, they are sorted sequentially according to the difficulty level of the learning passage (CT) that the learner studied immediately before, and then among the sorted learning passages, the learning passage that is first provided to the learner is provided in the order in which it is first exposed.

[0087] The provision of such recommended learning passages (CT) is provided after the learner has completed learning one learning passage (CT) through the learning confirmation unit (50) described later.

[0089] Meanwhile, the above CAT conversion unit (20) may be configured to generate a new image association character (CAT) for a new word when a new word is input or determined to be a new word in addition to the word stored in the database unit (10), and to transmit and store the new image association character (CAT) among the generated new image association characters (CAT) that has been verified by a worker who input the original text, such as a teacher or instructor, to the database unit (10).

[0090] Looking at this in more detail, it includes a new word verification unit (21) that checks whether there is a new word during the process of extracting words from the input original text, an abbreviation generation unit (22) that generates an abbreviation for the new word when it is confirmed as a new word by the new word verification unit (21), a verification unit (23) that verifies the generated abbreviation by the worker who input the original text, and a matching unit (24) that matches the image association character verified to be used as a new image association character with the new word confirmed by the new word verification unit (21) and transmits it to the database unit.

[0092] The above new word verification unit (21) identifies the word as a new word when there is no matching image association character from the database unit (10) in the step of extracting the original text input by the worker word by word and matching the extracted word with the image association character (CAT) stored in the database unit (10).

[0093] In addition, as another example, a separate new word input section (211) is provided and can be used when a worker increases the number of words stored in the database section (10).

[0094] When a new word is identified through the new word verification unit (21) above, an abbreviation is generated for the word. Since there may not be an image prepared for use as an image association character (CAT) for the new word, the abbreviation is generated first among the image and the abbreviation and provided to the image association character worker for the new word.

[0095] Here, if an image association character for a new word is prepared as an image, an upload window (not shown in the drawing) for uploading the image to be used as the image association character may be provided to the operator. This upload window is provided through the display of a terminal operated by the operator and is configured as a platform in which operations necessary for image uploading are controlled through input means such as a keyboard, mouse, and pen. Since the specific structure and operational relationship of such a platform are known, a detailed description will be omitted.

[0096] It is preferable for the abbreviation generation unit (22) to generate multiple abbreviations for a single new word and provide them to the worker, because the spelling to be emphasized may differ for each English word.

[0097] For example, if the new word is 'comprehensible', the abbreviations for the word are generated as 'compreh~le', 'compre~sibl', and 'comprehen~le', and then provided to the worker so that they can select the appropriate abbreviation.

[0098] The verification unit (23) provides a plurality of abbreviations generated through the abbreviation generation unit (22) to the worker for verification, and the worker selects one of the plurality of abbreviations provided through the verification unit (23) as an image association character for a new word.

[0099] The above matching unit (24) matches a new word and an abbreviation selected as an image association character for it through the above verification unit (23), and then transmits and stores it to the above database unit (10).

[0101] The next fingerprint unit (30) provides the learning fingerprint (CT) generated through the CAT conversion unit (20) to the learner online or offline. To this end, the fingerprint unit (30) is connected to an electrical communication line such as a wired or wireless communication means and transmits the learning fingerprint (CT) to an authenticated learner through a login server (not shown) or an authentication server (not shown).

[0102] Here, offline provision of the fingerprint study (30) means providing the learning fingerprint (CT) to multiple users through an external device, such as a beam projector installed in a classroom, for example.

[0104] The next learning provision unit (40) plays a recording file containing the original pronunciation of the original text that was previously saved for the original text by a native speaker or an instructor capable of providing the original text, after the learning text (CT) has been provided to the learner through the text provision unit (30) and after it has been confirmed that the learner has received the learning text (CT).

[0105] The above-mentioned learning provider (40) provides a learning text (CT) composed of image association characters (CAT) to a learner using the original text, which is the correct answer to the learning text (CT), with the original text recorded by a native speaker or an equivalent instructor. At this time, the recording file is provided repeatedly at regular intervals.

[0106] In other words, the learner infers and remembers the original words for the image association characters (CAT) that constitute the learning text (CT) by using sight to verify the learning text (CT) while listening to the audio file through hearing.

[0107] Through this, learners can improve memorization skills by matching images or abbreviations constituting Image Association Characters (CAT) to the original word, thereby facilitating long-term memory compared to conventional memorization by writing. Additionally, since learners can remember the original word by listening to its pronunciation along with the images or abbreviations of the Image Association Characters (CAT), they can simultaneously experience both image-based and auditory association methods, enabling them to remember the original word.

[0109] Next, the above-mentioned learning confirmation unit (50) receives the answer voice (AT) written by the learner regarding the learning text (CT) after the learner's learning through the above-mentioned learning provision unit (40) is completed, and determines whether it is correct.

[0111] The above learning confirmation unit (50) receives a voice answer sheet spoken by the learner in their own voice and determines whether it is correct. Here, the learner speaks the voice answer sheet for the learning text (CT) directly through a microphone installed on a smartphone, tablet, PC, etc., and the spoken voice file can be converted into text (Speech to Text) through a speech recognition program and then delivered to a teacher or instructor.

[0112] The learning verification unit (50) for this purpose includes a voice recognition unit (51) that recognizes a sound uttered by a learner, a word conversion unit (52) that converts the recognized sound into a word, a speaking verification unit (53) that converts the word determined to be the correct answer by comparing the converted word with the answer text and then provides it to the learner as a learning original word included in the original text, and a repetitive listening unit (54) that plays back a sentence containing a sound that is not converted into a word or a word determined to be an incorrect answer among the recognized sounds through the recording file.

[0113] The above-mentioned voice recognition unit (51) receives a sound spoken by the learner or a sound recorded by the learner, and the above-mentioned word conversion unit (52) converts the sound recognized by the above-mentioned voice recognition unit (51) into text.

[0114] The above-mentioned word conversion unit (52) uses a voice recognition program that converts a spoken or recorded sound into text. In the present invention, the voice recognition program classifies the learner's intonation, the loudness of the voice, i.e., the decibel of the sound, and the speed of speech to determine whether the answer is correct. In the process of converting the learner's voice into text, it prevents the word from being converted into a word different from the learner's intention due to the learner's intonation, or the spelling of the word from being changed due to the sound being too quiet to convert into a word, or the speed being too fast or too slow.

[0116] To this end, the word conversion unit (52) includes a voice correction unit (55) that corrects intonation, voice volume, speech speed, etc., which differ for each learner, and the voice correction unit (55) includes an intonation correction unit (551) that checks and corrects the learner's intonation, a decibel correction unit (552) that checks the voice volume, and a speed correction unit (553) that corrects the speech speed.

[0117] The above intonation correction unit (551) checks the pronunciation state of the learner during the process of converting the original word spoken by the learner into text and determines that the pronunciation is poorer when it is smaller than the standard value of 100% that is stored in advance, and the pronunciation is good when it is higher than the standard.

[0118] For example, if a learner's pronunciation is judged to be less than 50% of the standard value, the pronunciation part can be converted into the phrase "intonation is very poor" and provided to a teacher or instructor, and at the same time, the sentence can be displayed on the device used by the learner (smartphone, tablet, PC, etc.) to induce recurrence.

[0119] Here, the judgment stage of the intonation correction unit (551) is divided into multiple stages such as less than 50%, 50% to 70%, 71% to 90%, 91% to 100%, and 101% to 120%, and for each divided stage, sentences such as 'Intonation is very poor,' 'Intonation is poor,' 'Intonation is excellent,' 'It is at a native speaker level,' and 'It is at a very excellent level among native speakers' can be provided in order. The composition of the sentences provided here may vary depending on the age of the learner. In addition to sentences, any one of the emoticons including emojis may also be provided.

[0120] In addition, the decibel correction unit (552) determines the loudness of the learner's voice, that is, the loudness of the sound, and determines that the sound is quieter the smaller it is compared to the reference value of 100%, and the sound is louder the larger it is compared to the reference value.

[0121] Here, the decibel correction unit (552) can be configured to determine the tone of the learner's pronunciation in the case of a tonal language (Chinese, Thai, Vietnamese, etc.) where sound variation functions as a key element in distinguishing word meaning.

[0122] In the case of English, since there are emphasized pronunciations, or accents, it is possible to determine whether the learner has applied the accented part of the pronunciation to the original word where the accent should be emphasized.

[0123] In addition, while the accent of each English word is already determined, considering that one does not pronounce every word with an accent when constructing a sentence, it is advisable to check the accent only on words containing important information in the relevant learning text (CT).

[0124] For example, when 'How are you?' is provided in the study passage (CT), the word 'are' in the sentence is pronounced at a relatively higher pitch compared to other words, because the most important information in the sentence is in 'are'.

[0125] It is preferable that the settings for the areas requiring accent in these sentences be set in advance by the teacher or instructor during the process of inputting the original text in the CAT conversion unit (20).

[0126] The above speed correction unit (553) is intended to determine the learner's vocalization speed and determine that the smaller the value compared to 100%, the slower it is, and the larger the value, the faster it is, in order to induce an appropriate vocalization speed.

[0127] For example, when pronouncing multiple adjacent words in the process of pronouncing English words, the speed of pronunciation is fast, and the syllables of the preceding word and the following word overlap. In this case, all words with overlapping syllables are judged as incorrect.

[0128] The speech recognition program established as described above checks the learner's pronunciation, volume, and vocalization speed through each of the correction units, compares the confirmed speech with the original words constituting the original text, and if it is the correct answer, converts the corresponding word in the learning text (CT) that the learner is studying into the original word in the image association character (CAT) and provides it.

[0129] In other words, the word that the learner answered correctly is converted from the Image Association Character (CAT) back to the original word and provided, and this operation is repeated until the entire Image Association Character (CAT) constituting the learning text (CT) is converted back to the original word.

[0130] Through this, the learner performs repetitive learning until all of the learning passages (CT) are converted into original words, and in particular, can improve speaking ability through the learning confirmation unit (50).

[0131] The next speaking verification unit (53) compares the converted words with the above-mentioned correct answer text and converts the words determined to be correct into original learning words included in the original text and provides them to the learner. It determines the correct answer by comparing the words converted into text with the correct answer text, i.e., original learning words, and provides the correct answer to the learner after converting the words determined to be correct into original learning words, and provides the incorrect answers in the form of image association text.

[0132] The above-mentioned repeated listening unit (54) allows the learner to repeatedly play a recording file of a word judged as incorrect or a sentence containing a word judged as incorrect, thereby enabling the learner to learn repeatedly.

[0134] Next, the foreign language writing learning method (S1) utilizing image association characters and artificial intelligence according to the present invention comprises the steps of: matching original learning words with image association characters that match them and storing them in a database (S10); inputting a passage to be provided to a learner (S20); recording the inputted passage with an accent, pronunciation, and speed equivalent to that of a native speaker or an instructor to generate a learning recording file (S30); analyzing the original learning words constituting the inputted passage to generate a learning passage converted into matched image association characters and storing the inputted passage as a correct answer passage (S40); providing the generated learning passage to the learner online via a telecommunication line or in a printed state on paper via an output device (S50); playing the learning recording file after the learning passage has been provided to the learner (S60); and checking the answer voice spoken by the learner to determine whether the image association sentence is correct and scoring it (S70).

[0136] The above database storage step (S10) involves matching and storing a specific image or abbreviation with an original word selected as a learning word. Here, the database is managed by a database management system (DBMS), which is a system that manages data storage. Since the specific structure and operational relationship of such a database are well known, a detailed description will be omitted.

[0138] The step of inputting a learning text to be provided to the next learner (S20) is a step in which a teacher or instructor inputs a learning text to generate a learning text to be provided to the learner.

[0139] Here, a teacher or instructor inputs learning texts through an input device connected to the above database, wherein the input device may consist of a keyboard, mouse, or optical scanner, etc.

[0140] The above-mentioned optical scanner scans learning passages written by a teacher or instructor, saves them as digital files, and identifies and extracts words from the saved digital files, i.e., images. Here, the identification and extraction of words from images can be performed using a separate character recognition program (Optical Character Recognition: OCR); since such a character recognition program is a known one, detailed description and illustration will be omitted.

[0142] The step (S30) is to generate a recording file of the content of a learning text to be provided to the next learner, recorded by a native speaker of the language constituting the text or by an instructor or teacher capable of pronunciation, intonation, and speed equivalent thereto.

[0143] In other words, the teacher or instructor generates a learning audio file by having a native speaker or recording the learning passage provided to the learner directly, and simultaneously stores the learning passage so that it can be used as a criterion for determining the correct answer in the grading stage described later.

[0145] The next step (S40) is to generate a learning passage by calling image association characters that match the learning words constituting the input learning passage from the database and arranging them to have the same sentence structure as the learning passage, and to save the input learning passage as the correct answer passage.

[0146] A learning passage is generated to be provided to learners by retrieving image association characters from a database that match the learning words constituting the learning passage entered by a teacher or instructor, and arranging them to form the same sentence structure as the learning passage.

[0147] Once the generation of the learning passage is complete, the corresponding learning passage is identified as the correct answer and saved by matching it with the learning passage.

[0148] Here, the storage and retrieval of learning passages, answer passages, and learning passages are performed through the aforementioned database. Since the matching, storage, and retrieval of such data are identical to the operations of moving, storing, and loading general digital data, a detailed explanation thereof will be omitted.

[0150] Meanwhile, in the step of generating the learning passage (S40), if there is no matching image association character among the learning words constituting the learning passage, a new image association character for the corresponding learning word is generated and stored in the database.

[0151] To this end, the step of generating a learning text (S40) includes a new word verification step (S41) for verifying whether there is a new learning word among the learning original words extracted from the input learning text, an abbreviation generation step (S42) for generating an image association character consisting of an abbreviation for the learning original word identified as a new word, a verification step (S43) for verifying and obtaining approval for use of the generated new image association character from an administrator, and a matching step (S44) for matching the verified and approved new image association character to the new learning original word and storing it in the database.

[0153] The above new word verification step (S41) involves extracting the learning text input by the administrator word by word and, in the process of matching the extracted words with image association characters stored in the database, checking whether there are matching image association characters in the database, and identifying words that do not have matching image association characters as new words.

[0154] Here, a separate new word input step (S411) is provided so that an administrator can store image association characters matching the new word in the database.

[0155] The next abbreviation generation step (S42) is a step of generating an abbreviation for a word identified as a new word in the new word verification step (S41). Since there may be cases where an image to be used as an image association character is not prepared for a new word, an abbreviation is generated first among an image or an abbreviation to be used as an image association character, and provided to the administrator as an image association character for the identified new word.

[0156] Here, if an image association character for a new word is prepared as an image, an image upload step (S421) may be provided to the administrator so that the image can be used.

[0157] In addition, in the abbreviation generation step (S42), it is desirable to generate multiple abbreviations for a single new word and provide them to the worker, because the spelling to be emphasized may differ for each English word.

[0158] For example, if the new word is 'comprehensible', the abbreviations for the word are generated as 'compreh~le', 'compre~sibl', and 'comprehen~le', and then provided to the worker so that they can select the appropriate abbreviation.

[0159] The next verification step (S43) is a step in which the generated abbreviation is verified by an administrator, and approval for use is granted after verification.

[0160] In this verification step (S43), multiple generated abbreviations are each verified by the administrator, and the difficulty level for each abbreviation is set simultaneously with the verification.

[0161] Here, the difficulty level can be automatically set based on the number of letters omitted or the position of omission in the case of image association characters consisting of abbreviations. For example, the lowest difficulty level is one where the number of letters omitted is within the range of 1 to 3, the medium difficulty level is one where the number of letters omitted is 3 to 5, and the highest difficulty level is one where the number of letters omitted is 3 to 5, but the omitted positions are non-consecutive.

[0162] The next matching step (S44) is a step of verifying and approving through the verification step (S43), matching the new image association character with the difficulty level set with the corresponding new learning word, and then storing it in the database.

[0164] Meanwhile, in the step (S40) of generating the learning text, the generated learning text is classified into one of difficulty level, situation, or type, and then transmitted to and stored in the database unit. The learning text classified into one of difficulty level, situation, or type is provided to the learner according to the learner's request or the adjustment of the teacher or instructor. As learning progresses, learning texts of higher difficulty are provided, but learning texts of similar situations or types are provided so that interest in learning can be maintained.

[0165] For example, as shown in Fig. 4, if the learning text stored according to classification by type is related to food, the learning text shown in Fig. 4 is provided to a learner who has studied the food-related learning text as the next recommended learning text to study, thereby increasing learning efficiency by continuously studying food or the learning text related thereto.

[0166] Here, recommended learning passages refer to learning passages stored by category in the aforementioned database, multiple learning passages classified as the same type as the learning passage the learner studied immediately before, and among the multiple selected learning passages, sequentially sorted according to the difficulty level of the learning passage the learner studied immediately before, and then provided in an order in which the learning passage provided to the learner for the first time among the sorted learning passages is displayed first.

[0167] The provision of such recommended learning passages is provided after the learner has completed learning one learning passage through the scoring step (S70) described below.

[0168] Of course, in addition to the method of providing learning passages according to the aforementioned classification, they may also be provided in a random form without any special rules.

[0170] Meanwhile, in the step (S40) of generating the learning text, if the original learning word extracted from the learning text during the process of converting it into an image association word consists of two or more words combined and used differently from the original meaning of the words, an idiom verification step (S45) may be further included in which an idiom association word matching the idiom is applied to the learning text.

[0171] To this end, the above database separately stores idiom-specific mnemonic characters that match 100 idioms commonly used in English, and when a source word extracted from a learning text is determined to be an idiom, the idiom-specific mnemonic character that matches the idiom is applied first to the learning text.

[0172] As previously mentioned, idioms are combinations of two or more words and can be classified into preposition + noun, noun + preposition, idioms used in the form of verb + object + preposition using specific verbs and prepositions, idioms used in the form of verbs, and be verb + past passive voice. Depending on the classification, representative examples of these idioms are provided as separate mnemonic characters dedicated to idioms.

[0173] At this time, the teacher or instructor can choose to provide the word recognized as an idiom as an idiom-specific associative character or as a general image associative character.

[0174] In addition, for collocations that are fixed combinations of words in idioms and have the same meaning as combinations of individual words, you can choose to provide each word as an image association character or as an association character dedicated to the idiom.

[0175] As an example of the above idiom, 'free of charge' means 'free of charge', and 'free of charge' can be provided as image association characters designated as individual words 'fre', ' / ', and 'chag', or as another example, '0$' can be provided as an idiom-specific image association character created in image form.

[0177] In the next step (S50) of providing the generated learning text to the learner, the generated learning text is provided to the learner online or offline.

[0178] In this case, offline provision includes providing content in the form of printouts on paper via a printer or providing it through external devices, such as beam projectors installed in classrooms for multiple learners.

[0180] The next step of playing the recording file (S60) is a step of playing the recording file generated through the step of generating the recording file (S30) so that the learner can listen to it, and the number of times it is played can be adjusted according to the difficulty level of the learning text or selected by the learner.

[0181] In the step (S60) of playing the recording file, as described above, when the learning text is provided to the learner online, that is, in the case of a device capable of displaying the learning text through a display such as a smartphone, tablet, or desktop, the color, size, brightness, etc. of the learning text can be changed in conjunction with the playback of the recording file to improve learning efficiency.

[0183] The next scoring step (S70) is a step in which the learner listens to the learning text played back through the recording file, converts the learning text consisting of image association characters into original words, that is, original words for learning, and recognizes the answer voice spoken to determine the correct answer.

[0185] The scoring step (S70) includes an answer receiving step (S711) for receiving a file recorded using an electronic device capable of storing sound, such as a mobile phone or a recorder, of the answer voice spoken by the learner; a word extraction step (S712) for extracting words from the received image file; a speaking verification step (S713) for converting the extracted words into original words for learning after comparing them with the correct answer text and providing them to the learner; and a repeated listening step (S714) for playing a sentence containing words judged as incorrect through the recording file.

[0187] The above response receiving step (S711) is a step of receiving the response voice spoken by the learner, and may receive the voice file recorded by the learner online, or receive the sound spoken by the learner in real time and store it in a separate database.

[0188] Here, the database separately stores the audio responses transmitted or recorded by the learner and allows them to be reviewed upon the request of the learner or teacher, thereby enabling the verification of the learner's learning progress even after the learning is completed.

[0189] The above word extraction step (S712) extracts each word constituting the received response voice, and in this case, if the response voice is converted into a digital file, the words are extracted through a speech-to-text conversion program.

[0190] As such speech-to-text conversion programs are used in various fields, a detailed explanation of their specific structure and operational relationships will be omitted.

[0191] The extracted words are arranged in the same order as the sentence structure constituting the response speech, but characters that are not extracted as words are displayed in their recognized form or replaced with a separate 'unrecognized' indicator.

[0192] The next speaking verification step (S713) checks whether each character matches the correct answer based on its position relative to the learning text stored as the correct answer text, converts the word determined to be the correct answer into the original learning word, and provides it to the learner.

[0193] In the case of an answer voice that is not ultimately confirmed as 100% correct through this speaking verification step (S713), the learner is requested to submit a corrected answer voice, and if the learner submits a corrected answer voice, the answer voice is scored by repeating from the answer reception step (S711) or the word extraction step (S712).

[0194] The next repeated listening step (S714) is a step in which, if the answer audio is not 100% correct as a result of checking the answer audio in the speaking verification step (S713) above, the recording file is played again so that the learner can perform repeated learning on the learning text.

[0196] The above scoring step (S70) is a step for verifying the learner's speaking ability. Through this step, the learner can improve their foreign language speaking ability by visually checking a learning text composed of image association characters, simultaneously listening to a recording file, writing an answer voice through their own vocalization, and receiving a judgment on whether the written answer voice is correct.

[0198] Meanwhile, in the speaking verification step (S73) of the scoring step (S70) of the present invention, the words are determined by classifying the learner's intonation, the volume of vocalization, i.e., the sound decibels and the speed of speech for the received answer voice, respectively. In addition, a voice correction step (S74) is included to prevent the words from being converted into a word different from the learner's intention due to the learner's intonation, or the spelling of the word from being changed due to the sound being too quiet, making conversion into a word difficult, or the speed being too fast or too slow during the process of converting the learner's voice into text as described above.

[0200] The above voice correction step (S74) includes an intonation correction step (S741) for checking and correcting the learner's intonation, a decibel correction step (S742) for checking the volume of the voice, and a speed correction step (S743) for correcting the speed of speech.

[0201] The above intonation correction step (S741) checks the learner's pronunciation status during the process of converting the original word spoken by the learner into text, and determines that the pronunciation is poorer when the value is smaller than the standard value of 100% that is stored in advance, and the pronunciation is good when the value is higher than the standard.

[0202] For example, if a learner's pronunciation is judged to be less than 50% of the standard value, the pronunciation part can be converted into the phrase "intonation is very poor" and provided to a teacher or instructor, and at the same time, the sentence can be displayed on the device used by the learner (smartphone, tablet, PC, etc.) to induce recurrence.

[0203] Here, the judgment stage of the intonation correction stage (S741) is divided into multiple stages such as less than 50%, 50% to 70%, 71% to 90%, 91% to 100%, and 101% to 120%, and for each divided stage, sentences such as 'Intonation is very poor,' 'Intonation is poor,' 'Intonation is excellent,' 'Native speaker level,' and 'Very excellent level among native speakers' may be provided in order. The composition of the sentences provided here may vary depending on the age of the learner. In addition to sentences, any one of the emoticons including emojis may also be provided.

[0204] In addition, the decibel correction step (S742) determines the loudness of the learner's vocalization, that is, the loudness of the sound, and determines that the sound is quieter the smaller it is compared to the reference value of 100%, and the sound is louder the larger it is compared to the reference value.

[0205] Here, the above decibel correction step (S742) can be performed to determine the tone of the learner's pronunciation in the case of a tonal language (Chinese, Thai, Vietnamese, etc.) where sound variation functions as a key element for distinguishing word meanings.

[0206] In the case of English, since there are emphasized pronunciations, or accents, it is possible to determine whether the learner has applied the accented part of the pronunciation to the original word where the accent should be emphasized.

[0207] In addition, while the accent of each English word is already determined, considering that one does not pronounce every word with an accent when constructing a sentence, it is advisable to check the accent only on words containing important information in the relevant learning text.

[0208] For example, when 'How are you?' is provided in a study passage, the word 'are' in that sentence is pronounced at a relatively higher pitch compared to other words, because the most important information in that sentence is in 'are'.

[0209] It is preferable that the settings for the areas requiring accent in these sentences be set in advance by the teacher or instructor during the process of inputting the original text in the step (S40) of generating the learning text.

[0210] The above speed correction step (S743) is intended to determine the learner's vocalization speed and determine that the smaller the value compared to a reference value of 100%, the slower it is, and the larger the value, the faster it is, in order to induce an appropriate vocalization speed.

[0211] For example, when pronouncing multiple adjacent words in the process of pronouncing English words, the speed of pronunciation is fast, and the syllables of the preceding word and the following word overlap. In this case, all words with overlapping syllables are judged as incorrect.

[0212] Through the voice correction step (S74) performed as described above, the learner's pronunciation, volume, and vocalization speed are checked, and the checked voice is compared with the original learning word constituting the learning text. If it is the correct answer, the corresponding word in the learning text that the learner is studying is converted from an image association character to an original learning word and provided.

[0214] As described above, the present invention has been explained by specific details such as specific components, limited embodiments, and drawings; however, this is provided merely to aid in a more comprehensive understanding of the invention, and the invention is not limited to the above embodiments. A person skilled in the art can make various modifications and variations from this description.

[0216] Accordingly, the scope of the present invention should not be limited to the described embodiments, and all things equivalent to or having equivalent variations to the claims set forth below, as well as the claims set forth below, shall be considered to fall within the scope of the concept of the present invention. Explanation of the symbols

[0218] 1: Foreign language speaking learning system utilizing image association characters and artificial intelligence 10: Database Section 20 : CAT converter 21: New Word Verification Unit 22: Abbreviation Generation Unit 23 : Verification section 24 : Matching section 30: Studying passage questions 40 : Study materials 50 : Learning Confirmation Section 51: Writing Judgment Section 52: Speaking Judgment Section 53 : Voice correction unit 531 : Intonation Correction Unit 532 : Decibel Correction Unit 533 : Speed ​​correction unit

Claims

Claim 1 A database unit (10) storing image association characters (CAT) corresponding to one or more specific images or abbreviations for each learning word; a CAT conversion unit (20) that, when an original text composed of one or more original learning words to be provided to a learner is input, extracts the original learning words included in the input original text, calls the image association characters (CAT) corresponding to the extracted words from the database unit (10) and converts them into image association characters, generates a learning text (CT) composed of image association characters, classifies the generated learning text (CT) through one of the classification methods of difficulty, situation, or type, and then transmits it to the database unit (10) for storage; a text document unit (30) that provides the learning text (CT) generated through the CAT conversion unit (20) to the learner online or offline; and after the receipt of the learning text (CT) provided through the text document unit (30) is confirmed, plays a recording file in which the original text is recorded with the original pronunciation that was previously stored by a native speaker or an instructor capable of providing the corresponding intonation and pronunciation. A learning provision unit (40); a learning confirmation unit (50) that receives the answer voice (AT) generated in real-time or as a recording file of the sound the learner pronounces regarding the learning text (CT) after the learner's learning through the learning provision unit (40) is completed, and compares it with the original text stored previously to determine whether it is correct;The learning verification unit (50) includes a voice recognition unit (51) that recognizes a sound uttered by the learner or a sound recorded by the learner, a word conversion unit (52) that converts the recognized sound into words, a speaking verification unit (53) that converts the words determined to be correct answers by comparing the converted words with the correct answer text into learning original words included in the original text and provides them to the learner, and a repetitive listening unit (54) that plays back sentences containing sounds that are not converted into words or words determined to be incorrect answers through the recording file, and when converting words included in the input original text into corresponding image association characters (CAT), one or more image association characters (CAT) are included for each word in the conversion, and the converted image association characters (CAT) are provided to the learner according to a preset difficulty level, provided in order from lowest to highest difficulty or according to the difficulty level set by the user, and a new word verification unit (21) that checks whether there are new words in the process of extracting words from the input original text, and the new word A foreign language speaking learning system utilizing image association characters and artificial intelligence, further comprising: an abbreviation generation unit (22) that generates an abbreviation for a new word when it is identified as a new word in the verification unit (21); a verification unit (23) that verifies the generated abbreviation to a worker who inputs the original text; and a matching unit (24) that matches an image association character verified for use as a new image association character with a new word identified in the new word verification unit (21) and transmits the result to a database unit; wherein the word conversion unit (52) includes a voice correction unit (55), and the voice correction unit (55) includes an intonation correction unit (551) that verifies and corrects the learner's intonation, a decibel correction unit (552) that verifies the volume of the voice, and a speed correction unit (553) that corrects the speed of speech, and wherein the decibel correction unit (552) is characterized by determining the tone of the learner's vocalization in the case of a tonal language where sound fluctuation functions as an element for distinguishing the meaning of the word. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete

Images