Gender classification based on FeedForward Backpropagation neural network

被引：0

作者：

Azghadi, S. Mostafa Rahimi ^{[1
]}

Bonyadi, M. Reza ^{[1
]}

Shahhosseini, Hamed ^{[1
]}

机构：

[1] Shahid Beheshti Univ, Dept Elect & Comp Engn, Tehran, Iran

来源：

ARTIFICIAL INTELLIGENCE AND INNOVATIONS 2007: FROM THEORY TO APPLICATIONS | 2007年

关键词：

gender classifications; Backpropagation neural network; pitch features; Fast Fourier Transform;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Gender classification based on speech signal is an important task in variant fields such as content-based multimedia. In this paper we propose a novel and efficient method for gender classification based on neural network. In our work pitch feature of voice is used for classification between males and females. Our method is based on an MLP neural network. About 96 % of classification accuracy is obtained for 1 second speech segments.

引用

页码：299 / +

页数：3

共 9 条

[1] Voice-based gender identification in multimedia applications [J].

Harb, H ;

Chen, LM .

JOURNAL OF INTELLIGENT INFORMATION SYSTEMS, 2005, 24 (2-3) :179-198

[2]

Konig Y., 1992, IJCNN International Joint Conference on Neural Networks (Cat. No.92CH3114-6), P332, DOI 10.1109/IJCNN.1992.226966

[3]

KONIG Y, 1992, IJCNN, V7, P332

[4]

MARSTON D, 1998, ICASSP 98, V12, P357

[5]

MARSTON DF, 1998, IEEE P ICASSP, V1, P357

[6]

MARTLAND P, 1996, ICSLP 96, V4, P2231

[7]

PARRIS ES, P IEEE ICASSP, P685

[8]

QUAST H, 20022 UCSD I NEUR CO

[9]

RIVAROL V, 1996, ICSLP 96, V2, P1081

← 1 →