Document editing system using bidirectional digital braille conversion engine

Abstract

The present invention relates to a document editing system using a bidirectional digital braille conversion engine, which supports a visually impaired user to perform real-time conversion and editing between braille and text. The system comprises: a processor including an artificial intelligence inverse braille translation algorithm; a multi-language support braille input interface that receives a braille input from a user and converts the braille input into text data; document editing software that edits and stores the text data; a multi-output interface for providing the edited text data to the user as braille and voice outputs; and an integrated accessibility module that performs conversion between text and braille in real time according to an input of the user and supports compatibility with various digital platforms.

Description

Document editing system utilizing a bidirectional digital Braille conversion engine

[0001] The present invention relates to an AI-based bidirectional digital Braille conversion engine and a document editing system utilizing the same. More specifically, it is a technology that provides a function to convert text data into Braille and Braille into text for visually impaired users, and supports visually impaired users in easily editing and accessing digital documents through this conversion process.

[0002]

[0003] Braille conversion technology for the visually impaired primarily provides text information converted into Braille via Braille output devices (e.g., Braille displays). Most existing Braille conversion systems support unidirectional communication—that is, conversion only from text to Braille—which limits visually impaired users' ability to directly edit or modify documents. Furthermore, these systems often fail to adequately reflect the individual needs of users and frequently lack support for conversions across various languages ​​or contexts.

[0004] Despite advancements in modern Braille conversion technology, systems that support real-time bidirectional conversion and provide increasingly accurate and user-friendly conversions by learning user input through AI are very limited.

[0005]

[0006] The objective of the present invention is to provide a document editing system utilizing a bidirectional digital Braille conversion engine that supports bidirectional real-time conversion between text and Braille to improve upon the problems of conventional technology, thereby enabling users to modify and access documents in real time, provides accurate Braille conversion considering context by utilizing AI-based natural language processing technology, and ensures compatibility with various digital platforms.

[0007]

[0008] To achieve the above objective, according to one embodiment of the present invention, the system comprises: a processor including an artificial intelligence inverse braille algorithm; a multi-language support braille input interface that receives braille input from a user and converts it into text data; document editing software that edits and saves the converted text data; a multi-output interface that provides the user with the edited text data as braille and voice output; and an integrated accessibility module that performs conversion between text and braille in real time according to user input and supports compatibility with various digital platforms, wherein the artificial intelligence inverse braille algorithm may be a document editing system utilizing a bidirectional digital braille conversion engine that uses POS-based natural language processing technology to ensure accurate text conversion according to context and analyzes the meaning of input data to provide an appropriate response.

[0009] According to another embodiment, the braille input interface may be a document editing system utilizing a bidirectional digital braille conversion engine that includes a function to directly recognize braille from an image using a deep learning model that mutually converts braille and corresponding characters.

[0010] According to another embodiment, the document editing software may be a document editing system utilizing a bidirectional digital Braille conversion engine that includes an artificial intelligence-based recommendation system that allows the user to directly adjust the format, font, and layout of the document through a user interface, while automatically providing optimal editing suggestions based on the user's editing history.

[0011] According to another embodiment, the multi-output interface provides various output options that a user can select and provides information through a Braille display, voice output, or a combination thereof, and each output option may be a document editing system utilizing a bidirectional digital Braille conversion engine that includes artificial intelligence technology that learns the user's preferences and past usage patterns to automatically optimize.

[0012] According to another embodiment, the document editing system may utilize a bidirectional digital braille conversion engine that continuously learns user input data and updates the algorithm to provide personalized conversion results according to the input pattern of a processor user including the reverse braille conversion algorithm.

[0013] According to another embodiment, the integrated accessibility module may be a document editing system utilizing a bidirectional digital Braille conversion engine that includes a function enabling users to maintain the same working environment across various devices through cloud-based data synchronization.

[0014] According to another embodiment, the processor including the artificial intelligence inverse braille algorithm may be a document editing system utilizing a bidirectional digital braille conversion engine that includes a function to improve the conversion accuracy between braille and text by performing part-of-speech tagging that reflects linguistic characteristics such as particles attached after nouns and verbs, taking into account the agglutinative characteristics of the Korean language.

[0015]

[0016] The present invention provides the following effects through a document editing system utilizing a bidirectional digital Braille conversion engine.

[0017] By supporting bidirectional real-time conversion between text and Braille, users can modify and access documents in real time, thereby significantly improving the speed and efficiency of document work. This enables visually impaired users to access information more quickly and easily, thereby enhancing their independence in the digital world.

[0018] By utilizing AI-based natural language processing technology to provide accurate context-aware Braille conversion, it minimizes conversion errors and enhances the precision of meaning transmission. This contributes to reducing misunderstandings in educational materials, official documents, and everyday communication.

[0019] By providing a consistent user experience across various digital devices and platforms, it ensures that users enjoy the same accessibility in any environment. This promotes digital inclusion by enabling users of varying skill levels to easily adopt new technologies.

[0020] By learning users' input patterns and preferences to provide a customized interface, it enables a user experience that better suits individual user needs. This increases user satisfaction and maximizes the usefulness of the system.

[0021] These effects will contribute to promoting information accessibility and effective participation in the digital environment for visually impaired users, and ultimately help expand social, educational, and vocational opportunities for the visually impaired.

[0022]

[0023] FIG. 1 is a configuration diagram of a document editing system utilizing a bidirectional digital Braille conversion engine according to an embodiment of the present invention.

[0024] FIG. 2 is a flowchart illustrating the operation sequence of a document editing system utilizing a bidirectional digital Braille conversion engine according to an embodiment of the present invention.

[0025] Figure 3 is a Korean notation method using a 6-point Braille system.

[0026] FIG. 4 is a flowchart illustrating the process of an artificial intelligence inverse algorithm correcting and learning input data according to an embodiment of the present invention.

[0027]

[0028] To fully understand the structure and effects of the present invention, preferred embodiments of the present invention are described with reference to the attached drawings. However, the present invention is not limited to the embodiments disclosed below, but can be implemented in various forms and various modifications can be made. The description of the embodiments is provided merely to ensure that the disclosure of the present invention is complete and to efficiently explain the scope of the invention to those skilled in the art to which the present invention pertains.

[0029] Typos can be easily found on the bulletin board of the *Wide Village* website operated by the Korean Federation of the Blind; for instance, there were cases where 'daeta' was misspelled as 'deta' and 'jaeraegim' as 'jeregim'. Typos occur when using voice-based screen reader and document creation devices because it is difficult to accurately distinguish between characters with similar pronunciations. For instance, it is not easy to accurately distinguish and transcribe the differences between 'ae' and 'e', ​​or 'wae', 'oe', and 'we'.

[0030] Since the document editing system utilizing the bidirectional digital Braille conversion engine of the present invention enables visually impaired users to easily create documents using Braille, it will contribute to promoting information accessibility and effective participation in the digital environment, and ultimately help expand social, educational, and vocational opportunities for the visually impaired.

[0031] A document editing system utilizing a bidirectional digital Braille conversion engine according to an embodiment of the present invention includes a processor comprising an artificial intelligence reverse Braille algorithm, a Braille input interface, document editing software, a multi-output interface, and an integrated accessibility module.

[0032] FIG. 1 is a configuration diagram of an embodiment using the system of the present invention on a mobile device and a Braille input / output device. The system of the present invention can be used on various devices, such as tablets and desktops, in addition to mobile devices, and devices such as Braille input / output devices, speakers, and monitors can be connected via Bluetooth and used as an input / output interface.

[0033] The processor of the present invention operates by integrating an artificial intelligence reverse braille algorithm that enables real-time mutual conversion between braille and text for visually impaired users. Based on deep learning and natural language processing (NLP) technologies, this processor converts braille input into text and text input into braille, and provides more accurate and efficient conversion by learning the user's input patterns and context.

[0034] While converting from text to Braille presents no significant difficulties as it involves simply converting according to rules, converting from Braille to text presents challenges because it is often impossible to determine which character to convert to among characters using the same Braille notation without understanding the context. For example, referring to Fig. 3, which illustrates the Korean notation method in the 6-Braille system, it can be observed that the Braille notations for 'ㅖ' and 'ㅆ' are identical. Consequently, the Braille notations for "아예" and "았" are also the same, leading to difficulties in reverse Braille translation. The present invention provides more accurate and efficient conversion through reverse Braille translation that considers the user's input patterns and context using an artificial intelligence reverse Braille translation algorithm.

[0035] The processor, which includes an AI inverse braille conversion algorithm, features contextual analysis capabilities based on Part of Speech (POS) natural language processing. By grasping the context of the input text through POS tagging, semantic analysis, dependency parsing, morphological tokenization, language models, and context understanding, it goes beyond simple text conversion to perform braille conversion tailored to the meaning and grammatical structure of each sentence.

[0036] POS tagging is the process of assigning a part of speech to each word in a sentence. For example, in the sentence “The wind blows,” “wind” is tagged as a noun, and “to blow” is tagged as a verb. However, in the sentence “to have an affair,” “wind” is used with a different meaning; when used with “to have an affair,” in this context, “wind” is interpreted not as “blowing wind” but as a meaning related to a “romantic relationship.” Therefore, POS tagging determines the meaning appropriate to the context.

[0037] POS-based semantic analysis allows for the analysis of a word's meaning within a context once its basic role is determined through part-of-speech tagging. For example, the word "bank" can mean either "fruit of a tree" or "financial institution." In the context of "withdrawing money from a bank," it is interpreted as a financial institution, whereas in the sentence "stepping on a ginkgo tree while taking a walk," it is interpreted as the fruit of a tree. As such, POS-based semantic analysis can determine the precise meaning of polysemous words by considering the context.

[0038] Dependency parsing analyzes the structure of a sentence by identifying the relationships between words within it. For example, in the sentence “Cheolsu ate an apple,” the primary roles of the sentence can be distinguished as follows: “Cheolsu” as the subject, “apple” as the object, and “ate” as the verb. This allows for a deeper understanding of the sentence's meaning and enables correct contextual interpretation by identifying the relationships between sentence components.

[0039] Morphological tokenization is a method that separates input text into morphemes, the smallest meaningful units, and then applies the rules necessary for Braille conversion by analyzing the part of speech and contextual role of each morpheme. In particular, it efficiently processes particles and endings in agglutinative languages, as well as base forms and affixes in inflectional languages, and maximizes the efficiency of Braille representation by applying abbreviations and acronyms to repetitive words and phrases. This process contributes to preserving the meaning of the text while maintaining context, thereby making the Braille conversion results concise and accurate.

[0040] Pre-trained AI language models enable bidirectional understanding of context. For example, in the sentence “I read a book and drank coffee,” the system can understand the relationship between “read” and “drank” to identify that “book” is associated with “reading” and “coffee” is connected to “drinking.” This understanding of context enables more natural and accurate language processing.

[0041] Since Korean is an agglutinative language characterized by the attachment of particles, POS tagging requires identifying particles (“을,” “는,” “에”) attached to nouns or verbs. By reflecting these specific grammatical features of the Korean language, accurate POS tagging and contextual analysis are performed.

[0042] The processor, which includes an AI inverse Braille translation algorithm, supports deep learning-based multi-language support and includes a conversion model. By including a deep learning model trained on multi-language Braille standards, it converts text input in various languages ​​according to their respective Braille standards. Currently, it supports Braille standards for major languages ​​such as Korean, English, and Japanese, and can automatically adjust conversion standards based on the user's language environment. By applying the latest deep learning algorithms, it is possible to learn new conversions tailored to Braille standards whenever they are continuously updated, thereby reflecting the latest Braille rules.

[0043] Processors incorporating AI-powered Braille-inverse algorithms feature user-customized conversion learning capabilities. By continuously learning expressions and input patterns frequently used by users, they provide personalized conversion results. For instance, by learning how specific Braille input is typically converted into text, they offer more accurate conversions for similar inputs in the future. Additionally, they include correction functions that analyze user error tendencies to automatically correct incorrect input or reduce repetitive typos. This improves user input accuracy, and the learned data contributes to enhancing the personalized user experience.

[0044] FIG. 4 is a flowchart illustrating the process of an artificial intelligence inverse algorithm correcting and learning input data according to an embodiment of the present invention.

[0045] The data input stage is the stage of receiving data from the Braille input interface.

[0046] The tokenization and normalization stage of words or sentences through Korean corpus learning is a step in which the input data is analyzed by an AI back-reverse algorithm based on the corpus data it has learned. In this stage, the words constituting the sentence are tokenized (separated and structured), checked for spelling or syntactic errors, and normalized into a standard form. For example, duplicate spaces or incorrect spacing are corrected.

[0047] The typo detection step is a step that detects spelling errors or words that do not fit the context in the input data.

[0048] The Braille conversion candidate generation stage is the step of generating correct Braille conversion candidates based on detected typos and input data. In this process, POS-based NLP technology is used to analyze the context and generate the most suitable Braille conversion result for each word.

[0049] The language and context evaluation stage assesses whether the generated Braille conversion candidates are contextually appropriate through a language model. In this stage, the Braille and text conversion results are compared and reviewed by considering sentence structure and context to ensure that meaning is preserved.

[0050] The final word correction stage is where the optimal correction result is selected based on language and context evaluation. Incorrect words or phrases are corrected, and based on this, Braille and text are converted to each other.

[0051] The correction result output stage is the step of outputting the final corrected result in Braille or voice through the output interface. Users can check this result in real time.

[0052] The user correction confirmation step is a stage where, after the correction results are displayed to the user, the user can check for any parts requiring correction and request additional corrections. Accuracy can be improved by incorporating user feedback.

[0053] The training data update stage is the step in which feedback and correction history from the correction process are stored as training data. Through this, the system continuously learns the user's input patterns and error tendencies, enabling it to provide more sophisticated and personalized transformations for subsequent inputs.

[0054] This correction process improves the quality of input data and enhances user convenience by maximizing the accuracy of Braille and text conversion for the visually impaired.

[0055] A processor incorporating an AI-powered Braille-to-Text algorithm can provide results in the format most familiar to the user through real-time conversion and output optimization. It performs real-time conversion between Braille and text, enhancing operational efficiency by delivering conversion results immediately upon user input. Furthermore, the converted results are provided through various output methods, such as Braille displays or audio output, allowing users to instantly verify the content.

[0056] The processor, which includes an AI-powered braille-reverse algorithm, features system integration and compatibility. Designed for compatibility with various digital platforms, it guarantees consistent performance across diverse devices, including mobile phones, desktops, and tablets. This enables users to utilize the braille conversion system anytime and anywhere, regardless of their environment.

[0057] The Braille input interface of the present invention allows visually impaired users to input Braille and converts it into text data in real time. This interface supports Braille standards for various languages ​​and provides customized conversion to the user based on artificial intelligence and deep learning technologies. It is designed to recognize and convert Braille standards for multiple languages, including Korean, English, Japanese, and Chinese. Users can input Braille suitable for their language environment according to default settings, and the interface automatically applies conversion rules appropriate for that language. It performs conversion in accordance with the latest Braille standards and grammar, and if new Braille regulations are created, it reflects those standards through automatic updates, enabling users of various languages ​​to utilize the system consistently.

[0058] The document editing software of the present invention provides a function to edit converted text data in real time, allowing the user to freely modify the text after converting Braille into text. The user can edit the text at the sentence and word level, and can convert the text back into Braille for verification if necessary. The input text is automatically checked for grammar and analyzed for context to ensure the accuracy of the editing. Through this, the user can correct errors in real time and complete the final document in the desired form.

[0059] The document editing software of the present invention provides a function that allows the user to directly adjust the format and layout of a document. The user can adjust the text font, size, line spacing, paragraph alignment, etc., and each adjustment is saved in a format optimized for the convenience of visually impaired users. This function enhances document readability and editing efficiency, and in addition to visually supported layout settings, enables the user to perceive changes in the document format through voice feedback.

[0060] The document editing software of the present invention is equipped with an AI-based recommendation system that learns editing patterns frequently used by the user and recommends suitable editing options for the next input. It remembers formats, words, and phrases repeatedly used by the user and provides a function that automatically recommends them when similar input occurs. This function provides a user-customized editing environment, thereby increasing editing speed and offering a more intuitive editing experience. For example, if there are frequently used expressions, the corresponding words or phrases are automatically suggested whenever they are entered, which can shorten the input process.

[0061] The document editing software of the present invention supports Braille and text formats of various languages ​​and supports Braille and text conversion suitable for the grammar and characteristics of each language. In addition, the user can save the converted data in various formats such as hwp, txt, and doc, thereby maximizing the scalability and usability of the document.

[0062] The multi-output interface of the present invention provides convenient accessibility by outputting edited text data in Braille and voice forms. Text data edited by the user is converted into Braille in real time and displayed through a Braille display. The user can immediately check the text in Braille form during the editing process and receive rapid feedback through the Braille display.

[0063] The multi-output interface generates appropriate speech output by considering the meaning and context of the text entered by the user through AI-based natural language processing technology. It minimizes pronunciation errors and provides natural voice feedback by automatically adjusting intonation and stress according to sentence structure. Speed, volume, and tone can be adjusted according to user preferences, and it provides the most suitable voice feedback by learning from previous feedback data. For example, specific phrases or words are emphasized according to the user's previous settings, or output at the desired speed, providing a convenient listening experience.

[0064] The multi-output interface allows for the individual use of Braille and voice outputs, as well as supports a combined output mode that utilizes both simultaneously. By providing both Braille and voice outputs, it helps users understand text information more clearly. It delivers information optimized to the user's preferred output method. For instance, users can easily switch output modes according to the situation, such as preferring Braille indoors while selecting voice output when on the move. This feature enhances user convenience and ensures appropriate accessibility in various situations.

[0065] The multi-output interface is compatible with various devices and provides Braille and voice output optimized for each device. This allows users to experience the same level of accessibility regardless of the specific device. With the voice output function available for use on the go, users can conveniently check edited text information regardless of their work environment.

[0066] The integrated accessibility module of the present invention is a component designed to provide accessibility and usability across various digital platforms. By supporting compatibility with diverse digital platforms such as mobile phones, tablets, and desktops, it offers a consistent user experience that allows users to access content regardless of the specific device or environment. This module provides an interface optimized for the characteristics of each platform, helping to maintain consistent usability across platforms. Consequently, users can experience the same level of Braille and text conversion functionality on any device.

[0067] The integrated accessibility module includes Braille output, voice output, and screen reading functions, allowing users to select the appropriate approach depending on the situation. This enables users to choose the method that suits them, such as preferring voice feedback while on the move or Braille output in static environments. Each approach operates independently, and for user convenience, a combined mode is also provided for simultaneous use, offering optimal feedback tailored to the task at hand.

[0068] The integrated accessibility module supports cloud-based data synchronization, allowing users to easily access and use documents or settings across various devices. This enables users to maintain workflow continuity and retrieve their previous work to continue editing anytime, anywhere. Cloud integration securely stores user settings and editing data, maximizing digital accessibility by providing the same editing environment on different devices.

[0069] The integrated accessibility module utilizes enhanced security protocols to safely protect user data. While AI-based accessibility features learn user input patterns and data, this information is securely managed in an encrypted state, prioritizing privacy protection. When using cloud integration and data synchronization features, security measures are provided to ensure user information is stored securely and prevents external leakage.

[0070] FIG. 2 is a flowchart illustrating the operation sequence of a document editing system utilizing a bidirectional digital Braille conversion engine according to an embodiment of the present invention. Specifically, the document editing system utilizing a bidirectional digital Braille conversion engine according to an embodiment of the present invention operates in the following steps: receiving Braille input (S1), converting data from Braille to text (S2), editing and modifying text data (S3), setting user-customized environments (S4), outputting Braille and voice (S5), storing in the cloud and synchronizing data (S6), and providing real-time feedback and error review (S7).

[0071] The Braille input reception step (S1) is a step of receiving Braille data through a Braille input interface. The user inputs Braille data that supports multiple languages ​​through the Braille input interface. This interface recognizes Braille standards of various languages ​​and transmits the input Braille data to the processor.

[0072] The step of converting data from Braille to text (S2) is a step in which a processor including an artificial intelligence inverse Braille algorithm converts Braille data received through a Braille input interface into text data. The processor including the artificial intelligence inverse Braille algorithm learns the context and the user's input pattern to improve the accuracy of the conversion, and generates text data in real time and transmits it to document editing software.

[0073] The text data editing and modification step (S3) is a step in which text data received from the processor is edited and modified in document editing software. The document editing software displays the text data in a document format and allows the user to freely modify the text.

[0074] The user-customized environment setting step (S4) is a step that provides a user-customized editing environment through an integrated accessibility module. The integrated accessibility module customizes the editing environment by learning the user's previous usage patterns and preferences, such as providing an optimized interface and feedback that considers the user's preferred output mode (Braille or voice) and device-specific accessibility.

[0075] The Braille and voice output step (S5) is a step of outputting edited text data in Braille and voice through a multi-output interface. The edited text data is output in Braille and voice forms through the multi-output interface. The user can check the final result using a Braille output display and a voice feedback function. The user can select the output method depending on the situation or utilize both outputs simultaneously.

[0076] The cloud storage and data synchronization step (S6) is a step of storing and synchronizing edit data in the cloud through the integrated accessibility module. The document that has been edited is stored in the cloud through the integrated accessibility module, and the data is synchronized so that it can be accessed from other devices. Through this, the user can retrieve the document they have worked on and continue editing it regardless of time and place.

[0077] The real-time feedback and error review step (S7) is a step for finally verifying the edited data through real-time feedback and error detection. The integrated accessibility module and the multi-output interface provide real-time feedback on the edited text data, allowing for the review and correction of errors in Braille and text conversion and output. If correction is necessary, the process is repeated starting from the text data editing and correction step (S3).

[0078] Once final editing is complete, the user-customized environment is maintained, and if necessary, it can be saved in other formats (hwp, txt, doc, etc.) for use.

Claims

1. As a document editing system utilizing a bidirectional digital Braille conversion engine, A processor including an artificial intelligence inverse algorithm; A multi-language supported Braille input interface that receives Braille input from a user and converts it into text data; Document editing software for editing and saving converted text data; A multi-output interface that provides edited text data to the user as Braille and voice output; and It is configured to include an integrated accessibility module that performs real-time conversion between text and Braille based on user input and supports compatibility with various digital platforms, and The above artificial intelligence reverse braille algorithm is a document editing system utilizing a bidirectional digital braille conversion engine that uses POS-based natural language processing technology to ensure accurate text conversion according to context and analyzes the meaning of input data to provide an appropriate response.

2. In Paragraph 1, The above-described Braille input interface is a document editing system utilizing a bidirectional digital Braille conversion engine that includes a function to directly recognize Braille from an image using a deep learning model that mutually converts Braille and corresponding characters.

3. In Paragraph 1, The above document editing software is a document editing system utilizing a bidirectional digital Braille conversion engine that includes an AI-based recommendation system that automatically provides optimal editing suggestions based on the user's editing history, while allowing the user to directly adjust the document format, font, and layout through a user interface.

4. In Paragraph 1, The above-described multi-output interface provides various output options selectable by the user and provides information through Braille display, voice output, or a combination thereof, and is a document editing system utilizing a bidirectional digital Braille conversion engine that includes artificial intelligence technology that automatically optimizes each output option by learning the user's preferences and past usage patterns.

5. In Paragraph 1, A document editing system utilizing a bidirectional digital Braille conversion engine that continuously learns user input data and updates the algorithm to provide personalized conversion results according to the input patterns of a processor user including the above-mentioned reverse Braille algorithm.

6. In Paragraph 1, The above integrated accessibility module is a document editing system utilizing a bidirectional digital Braille conversion engine that includes a function enabling users to maintain the same working environment across various devices through cloud-based data synchronization.

7. In Paragraph 1, A document editing system utilizing a bidirectional digital Braille conversion engine that includes a processor including the above-mentioned artificial intelligence inverse Braille conversion algorithm, performs part-of-speech tagging reflecting linguistic characteristics such as particles attached after nouns and verbs by considering the agglutinative characteristics of the Korean language, and thereby improves the accuracy of conversion between Braille and text.

Images