TY  - JOUR
T1  - A Study on the Voice Authentication Security for Recorded Voice through
Sound Color Marker Analysis
AU - Bae, Myung-Jin AU - Kim, Bong-Young 
JO  - Journal of Engineering and Applied Sciences
VL  - 14
IS  - 21
SP  - 7922
EP  - 7927
PY  - 2019
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2019.7922.7927
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2019.7922.7927
KW  - Sound color marker
KW  -voice authentication
KW  -biometric authentication
KW  -security
KW  -recorded voice
KW  -authentication
AB  - With the development of technology related to the 4th Industrial Revolution such as smart phone
and internet of the things, more convenient financial services with biometric authentication are emerging one
after another. Voice authentication is cheaper and easier to introduce than other biometrics authentication
methods because it can provide the biometrics authentication function with only the microphone installed in
the smart phone. Voice authentication is a very good authentication method because it is possible to apply the
speaker presentation method such as sentence presentation in terms of security. Especially, since, it is very
difficult to restore human voice to a common speaker, speech recognition is a highly secure biometric
authentication method. In this study, we used sound color marker that express intuitively the voice components
to see, if it is possible to distinguish between real voice and recorded voice. As a result of the experiment in
Chapter 3, we could distinguish the real voice from the recorded voice by analyzing the similarity of the sound
color marking and it was confirmed that the voice is very secure from the recorded voice and the processed
voice.
ER  - 