|National Conference on Recent Trends in Information Technology
|Foundation of Computer Science USA
|NCIT2015 - Number 1
|Authors: Seema Khanum, Marpe Sora
Seema Khanum, Marpe Sora . Speech based Gender Identification using Feed Forward Neural Networks. National Conference on Recent Trends in Information Technology. NCIT2015, 1 (January 2016), 5-8.
This paper proposes an efficient method of gender identification based on the speaker's voice in a noisy environment. MFCC was used to extract features from the speech sample taken from a noisy speech database; these features are then used to train Artificial Neural Network architecture to classify two different genders (Male and Female). The test result shows that the new proposed ANN architecture can analyze and learn better and faster. The advantage of proposed method is a result of decreasing the number of segments by grouping similar segments in training data using a clustering technique namely fuzzy c means clustering.