Biometric Person Identification using Multilingual Speech Processing |
Author(s): |
| Tushar Kant Sahu , SHRI SHANKRACHARYA COLLEGE OF ENGINEERING AND TECHNOLOGY BHILAI; Vinay Jain, SHRI SHANKRACHARYA COLLEGE OF ENGINEERING AND TECHNOLOGY BHILAI |
Keywords: |
| Multilingual Speaker Identification, MFCC, Speaker Identification, Support Vector Machine, Indian Languages |
Abstract |
|
In this paper we explain the multilingual speaker identification system. Speaker identification is conducted on 3 Indian languages Hindi, Marathi and Rajasthani. We create a database of 15 person in each language. In our system we use 3 different sentences and each sentence in 3 languages. We focus on the effect of language mismatch in the speaker identification performance of individual languages and all languages together. Mel Frequency Cepstral Coefficient (MFCC) is used for feature extraction. The standard SVM-based speaker identification is used. The average speaker identification rate is as high as 99%. |
Other Details |
|
Paper ID: IJSRDV4I100546 Published in: Volume : 4, Issue : 10 Publication Date: 01/01/2017 Page(s): 687-691 |
Article Preview |
|
|
|
|
