BoryVoice Solution

BoryVoice Solution | Bory Speech Recognition and Lip-Reading STT Solution |

An AI STT solution that analyzes speech and lip-movement video together to convert conversations into text and extend them to real-time captions, consultation records, and accessibility services

BoryVoice Solution is a speech recognition and lip-reading STT solution that implements speech recognition, lip-reading AI, real-time STT, caption generation, conversation recording, and audio-video speech analysis functions tailored to the customer’s service environment.

In everyday conversations, public service consultations, hospital reception, meetings, lectures, kiosk ordering, AR glasses captions, and customer center consultations, it is important to convert speech into text accurately. However, in real-world environments, speech recognition alone is often insufficient due to background noise, speaker distance, mask wearing, differences in pronunciation, multiple people speaking at the same time, and poor audio input quality.

BoryVoice Solution uses both audio information and lip-movement video to convert conversations into text and can integrate the results with real-time captions, consultation records, conversation summaries, accessible displays, AR glasses, kiosk guidance, and chatbot or consultation systems.

This AI speech and lip-reading STT solution can be applied to public institutions, hospitals, welfare organizations, educational institutions, corporate customer centers, AR glasses services, kiosks, robots, video platforms, hearing-assistance services, and barrier-free consultation systems.

 

[Inquire About Solution Deployment]

[Consultation on Building Speech Recognition and Lip-Reading STT]

[PoC Development Inquiry]

[Request a Proposal]

Customer Challenges

 

Speech recognition alone cannot reliably process on-site conversations

In actual service counters, stores, hospitals, public institutions, classrooms, and meeting rooms, speech recognition results can fluctuate because of noise, distance, speaking habits, microphone quality, and other people’s voices.

BoryVoice Solution supports caption generation and STT quality by using a structure that combines audio information with lip-movement video.

 

Real-time captions are needed for people who are deaf or hard of hearing

People who are deaf or hard of hearing, as well as older adults, may have difficulty hearing face-to-face consultations or spoken guidance.

BoryVoice Solution converts conversations into text in real time and displays them on caption monitors, AR glasses, smartphones, tablets, or web screens to improve hearing accessibility.

It is difficult to record and manage consultation content

Public service consultations, hospital reception, customer centers, educational consultations, and product consultations require the content to be recorded, summarized, and managed for follow-up actions.

BoryVoice Solution converts audio- and video-based conversations into text and integrates them with consultation records, summaries, search, reports, and agent review workflows.

 

Kiosks and robots need natural voice interfaces

Touchscreens alone may not be convenient for every kiosk or robot user.

BoryVoice Solution can connect voice commands, voice ordering, spoken guidance, and conversational interfaces so customers can place orders or receive guidance by speaking.

 

Speech in video needs to be captioned

Lecture videos, consultation videos, guidance videos, educational content, and meeting videos are easier to use for search, summarization, accessibility support, and learning assistance when captions are available.

BoryVoice Solution uses speech and lip-movement information in video and can be extended to video caption generation, conversation recording, speech-segment analysis, and summarization.

 

Solution Overview

BoryVoice Solution is an implementation-oriented STT solution that converts audio- and video-based conversations into text and integrates the results with various services.

 

Key features include the following.

• Speech Recognition-Based STT
• Provide lip-reading-AI-based caption assistance
• Audio-and-Lip Combined STT
• Real-Time Conversation Caption Generation
• Video File-Based Caption Generation
• Consultation Content Transcription
• Conversation Record Storage
• Speaker Diarization Integration
• Keyword Extraction
• Conversation Summarization Integration
• Easy-to-Understand Sentence Conversion Integration
• AR Glasses Caption Display Integration
• Kiosk Voice Ordering Integration
• Chatbot and Consultation System Integration
• Administrator Dashboard Development
• External Service API Integration

 

Key Features

 

1. Speech Recognition-Based STT

BoryVoice Solution implements STT functionality that converts speech captured through a microphone into text.

Various types of audio data, including consultations, meetings, lectures, guidance, orders, and customer service interactions, can be transcribed in real time or from files.

 

2. Lip-Reading AI-Based Caption Assistance

When speech alone is difficult to recognize, lip-movement video information can be used as supplementary input.

The solution can be extended into a lip-reading STT structure that analyzes an agent’s lip movements, the speaker’s facial area, and speech segments to supplement speech recognition results.

 

3. Audio-and-Lip Combined STT

BoryVoice Solution can be configured as an audio-visual STT structure that uses both audio and video data.

It can be used as a caption assistance function in environments where speech recognition quality decreases because of heavy noise, unclear pronunciation, mask wearing, distance, or similar issues.

 

4. Real-Time Caption Generation

Real-time captions can be generated during consultations, meetings, lectures, public service guidance, hospital reception, and kiosk interactions.

Captions can be displayed on monitors, web screens, mobile screens, AR glasses, or kiosk screens, with text size, position, display speed, and high-contrast mode adjusted for the usage environment.

 

5. Video Caption Generation

Speech can be extracted from recorded video files or live video streams to generate captions.

It can be used for educational videos, consultation videos, meeting videos, guidance videos, promotional videos, and accessible content production.

 

6. Consultation Record Automation

Customer consultations or public service consultations can be converted into text and stored as consultation histories.

Agents can review and edit AI-generated records and then use them for internal records, consultation summaries, follow-up actions, and customer guidance materials.

 

7. Speaker Diarization Integration

Speaker diarization can be integrated when multiple people are conversing.

By distinguishing who said what in meetings, consultations, education, and customer service, the usefulness of conversation records and summaries can be improved.

 

8. Conversation Summarization and Keyword Extraction Integration

BoryVoice Solution can connect STT results with BoryTalk Solution or BORY Sense Solution and extend them to conversation summarization, key keyword extraction, and follow-up action organization.

Long consultations or meetings can be organized around key sentences, major questions, answers, decisions, and requests.

 

9. Kiosk Voice Ordering Integration

By integrating with BoryKiosk Solution, customers can search for menu items and place orders by voice.

For example, utterances such as “One Americano,” “Make it iced,” “Use my points,” or “I’ll pay now” can be connected to the ordering flow.

 

10. AR Glasses Caption Integration

When integrated with BoryGlass Solution, conversations can be displayed as real-time captions on AR glasses.

Users can view captions while looking at the other person, enabling hearing assistance and conversational accessibility services.

 

Available Deployment Configurations

 

Basic Deployment Configuration

• Audio Input Module
• STT Conversion Module
• Real-Time Caption Generation Module
• Caption Display Screen
• Conversation Record Storage Module
• Administrator Screen
• Results Viewing Screen
• API Integration Module

 

Extended Deployment Configuration

• Lip-Reading AI Analysis Module
• Audio-and-Lip Combined STT Module
• Video File Caption Generation Function
• Live Video Streaming Integration
• Speaker Diarization Function
• Keyword Extraction Function
• Conversation Summarization Function
• Plain-language conversion function
• AR Glasses Caption Display Integration
• Kiosk Voice Ordering Integration
• Consultation System Integration
• Customer Center System Integration
• Robot Conversation Engine Integration
• Cloud server deployment
• On-premises server deployment
• Deployment on the Institution's Internal Network
• Customer Service API Integration

 

Application Scenarios

 

Public Institution Consultation Captions

Real-time captions for staff speech can be provided at public institutions, local governments, community service centers, and administrative welfare centers.

When integrated with BoryGlass Gov Solution, it can be implemented as a barrier-free AI public service consultation system for people who are deaf or hard of hearing.

 

Hospital Reception and Consultation STT

At hospital administration offices, reception counters, and examination guidance counters, reception procedures, examination guidance, payment guidance, and medical service information can be provided as text.

In hospital environments with many older adults and people who are hard of hearing, viewing consultation content on screen can help reduce communication errors.

 

AR Glasses Hearing-Assistance Captions

In conjunction with BoryGlass, everyday conversations, meetings, lectures, and consultations can be displayed as captions on AR glasses.

It can be used to improve conversational accessibility for people who are deaf or hard of hearing and older users.

 

Kiosk Voice Ordering

At restaurants, cafés, unmanned stores, hospitals, and public institution kiosks, customers can select menu items or receive guidance by voice.

When linked with BoryKiosk Solution, a barrier-free kiosk environment combining voice ordering, caption guidance, and an easy-to-use UI can be implemented.

 

Meeting and Lecture Captions

Speech can be converted into captions in meeting rooms, classrooms, seminars, and online education environments.

It can be used for meeting minutes, lecture captions, learning assistance, and stronger hearing accessibility.

 

Customer Center Consultation Records

In customer centers, call centers, online consultations, and video consultations, consultation content can be converted into text and summarized.

It can be used for consultation quality management, consultation history searches, complaint response, and training material production.

 

Video Content Captioning

Automatic captions can be generated for educational videos, promotional videos, guidance videos, consultation videos, and meeting videos.

When integrated with BORY Sense Solution, it can be extended beyond captions to summarization, plain-language explanations, and video understanding functions.

 

Robot Conversation Interface

Speech recognition functionality can be integrated with YAONG, YOYO, guide robots, care robots, and educational robots.

The system recognizes the user’s speech and connects it with the BoryTalk conversation engine so the robot can provide guidance, companionship, consultation, and educational conversations.

 

Expected Benefits

 

Improved Conversation Accessibility

By converting spoken conversations into text, the solution helps people who are deaf or hard of hearing, older adults, and foreign-language users understand information more easily.

 

Improved Consultation Efficiency

Automatically transcribing and summarizing consultations reduces the burden of manual record keeping and allows follow-up work to be handled more quickly.

 

Improved Kiosk and Robot Usability

Voice-based commands and guidance can be provided for users who have difficulty operating touchscreens or buttons.

 

Greater Use of Video Content

Converting speech in video into captions and text enables search, summarization, education, and accessible content production.

 

Enhancement of Existing Services

Speech recognition and lip-reading STT functions can be added to websites, apps, kiosks, AR glasses, consultation systems, customer centers, and robot services.

 

Connection with Bory Accessibility Solutions

BoryVoice Solution can connect with BoryGlass, BoryGlass Gov, BORY Sense, BoryKiosk, and BoryTalk and serve as the core voice and caption engine for barrier-free consultation, hearing assistance, and AI customer service.

 

Deployment Process

 

1. Requirements Review

The customer’s intended use, target users, audio input environment, availability of video input, caption display method, and systems to be integrated are reviewed.

 

2. Data Environment Analysis

Microphone quality, camera position, lighting, noise, speaker distance, spoken language, video quality, and network environment are analyzed.

 

3. Functional Design

Required functions are defined, including speech recognition, lip-reading assistance, real-time captions, video captions, consultation records, kiosk voice ordering, and AR glasses display.

 

4. PoC Implementation

For selected environments or speech data, STT quality, caption display, lip-reading assistance, screen integration, and API integration feasibility are verified.

 

5. Full Implementation

An STT server, lip-reading analysis module, caption screen, administrator screen, conversation records, and API integration functions are implemented for the customer’s environment.

 

6. Operational Verification and Enhancement

Recognition accuracy, caption quality, screen readability, summarization quality, and user feedback are reviewed and improved based on actual usage data.

 

What BoryVoice Solution Can Do

 

• Build a Speech Recognition STT System
• Provide lip-reading-AI-based caption assistance
• Build an Audio-and-Lip Combined STT System
• Real-Time Conversation Caption Generation
• Generate Captions from Video Files
• Consultation Content Transcription
• Build a Meeting and Lecture Caption System
• Integrate Public Institution Consultation Captions
• Integrate Hospital Reception Counter Captions
• AR Glasses Caption Display Integration
• Kiosk Voice Ordering Integration
• Integrate Robot Voice Conversation
• Automate Customer Center Consultation Records
• Integrate Conversation Summarization
• Integrate Easy-to-Understand Sentence Conversion
• Build an STT API
• Build an On-Premises Speech Recognition System
• Build a Cloud-Based Speech and Caption Service

 

Implementation Inquiry

 

Depending on the customer’s usage environment, audio and video data structure, caption display method, integrated systems, and security policy, BoryVoice Solution can be proposed as a PoC, STT API implementation, real-time caption system, kiosk voice ordering integration, AR glasses caption integration, or consultation record automation project.

• Inquiry About Adopting the Speech Recognition and Lip-Reading STT Solution
• Inquiry About Building a Real-Time Caption System
• Lip-Reading AI PoC Inquiry
• Inquiry About Building Public Institution Consultation Captions
• Inquiry About a Hospital Reception Counter Caption System
• AR Glasses Caption Integration Inquiry
• Kiosk Voice Ordering Integration Inquiry
• Inquiry About Building a Meeting and Lecture Caption System
• Customer Center Consultation Record Automation Inquiry
• STT API Development Inquiry
• Request a Proposal

 

Important Information

BoryVoice Solution is an AI-based STT solution that uses speech and lip-movement video information to convert conversations into text and support captions, consultation records, and accessibility services.

This solution assists communication and information understanding and does not guarantee complete recognition of every spoken utterance or lip movement.

 

Speech recognition and lip-reading STT results may vary depending on background noise, lighting, camera position, microphone quality, speaker distance, speaking habits, mask wearing, obstruction of the mouth, video quality, and network conditions.

If consultation audio, video, captions, conversation records, or user information are stored or transmitted, prior consultation regarding privacy standards and data security policies is required.

 

For consultations requiring important judgments, such as medical, legal, financial, or public service matters, it is recommended that a responsible staff member review AI-generated STT results and summaries.

 

Before adoption, the intended use, audio and video collection methods, caption display method, data storage scope, personal information processing standards, and integrated systems must be reviewed in advance.

 

---

BORY.ai Key Services Overview

Bory Co., Ltd. develops products, solutions, platforms, and services for industrial, medical, public-sector, healthcare, and barrier-free applications based on AI technologies involving speech, language, video, sensors, and data.

Below are the main representative domains currently operated or being prepared by Bory Co., Ltd. 

 

Primary Domain	Service/Brand	Description
bory.ai	BORY.ai	The official AI brand website of Bory Co., Ltd., serving as the company’s main website for the integrated presentation of its products, solutions, platforms, and services
borysense.com	BORY SENSE	An AI-powered hearing assistance platform that supports communication for people with hearing disabilities and older adults through real-time captioning, lip-reading AI, and AR glasses integration
borytalk.com	BORY TALK	A web-based conversational AI chatbot service designed for civil service inquiries, consultations, information guidance, and customer support
borykiosk.com	BORY KIOSK	A barrier-free AI kiosk service for older adults and people with disabilities, featuring voice guidance, captioning, and easy-to-use interfaces
boryservice.com	BORY SERVICE	A service portal introducing Bory’s AI services and custom-built AI offerings for industrial, public-sector, and everyday applications
borysong.com	BORY SONG	A music AI service that supports AI-powered composition, music generation, and sound content production