Voice ID

 

Introduction

The Voice ID (VID, speaker recognition) service is a technology that analyzes the unique characteristics of a user's voice to identify the user.

The process for identifying a speaker in the Voice ID service is as follows:

1. Pre-learning

   1. Learns a deep learning model based on speaker voice information of large-capacity.

2. Speaker registration

   1. Voice data collection

      • Receives voice data from users and collects voice data samples to be used for speaker recognition. 

   2. Feature extraction

      • Extracts unique features from collected voice data. In this process, information such as the user's voice characteristics, pronunciation, voice pitch, and stress is analyzed.

   3. Model training

      • A machine learning algorithm is used to learn an individual speaker recognition model based on the extracted features.

3. Speaker identification

   1. When a user utters, the speaker is identified based on information registered in the dictionary.

 

The Voice ID service provided by LG AI Platform has the following features:

ASR 엔진의 특징

Features	Description
High accuracy	The speaker identification accuracy of the Voice ID service is high. The voice data is collected six times to increase accuracy. Also, the data collected in various environments has higher accuracy.

 

Structure

The Voice ID service receives voice data as input value, extracts the data, and provides the result to the application.

 

Examples of Use

The Voice ID can be used in a variety of services that require speaker recognition for more diverse and efficient interactions with users.

Smart home system

The system recognizes a user who uttered to provide personalized services suitable for the user.

Meeting minutes preparation

The system can provide a service that identifies the utterer and prepares meeting minutes.