VOICE SIGNAL RECOGNITION SYSTEM RECOGNIZING VOICE SIGNAL OF EACH OF MULTIPLE USERS USING PERSONALIZED LAYERS CORRESPONDING TO EACH OF MULTIPLE USERS

摘要

A voice signal recognition system may minimize the difference in voice recognition performance between users using personalized layers stored for each device and learning the utterance characteristics of a user. Since personalized layers are not stored in a voice signal recognition server to which devices are connected, costs of creating and storing the personalized layers corresponding to each of the multiple users using the voice signal recognition system may be reduced.;COPYRIGHT KIPO 2019