Emotion Recognition using combination of MFCC and LPCCwith Supply Vector Machine
Journal Title: IOSR Journals (IOSR Journal of Computer Engineering) - Year 2015, Vol 17, Issue 4
Abstract
Abstract: Speech is a medium through which emotions are expressed by human being. In this paper, a mixtureof MFCC and LPCC has been proposed for audio feature extraction. One of the greatest advantage of MFCC isthat it is capable of identifying features even in the existence of noise and henceforth it is combined with theadvantage of LPCC which helps in extracting features in low acoustics. Two databases have been considerednamely Berlin Emotional Database and a SAVEE database. SVM has been implemented for classification ofseven emotions (Sadness, Joy, Fear, Anger, Boredom, Disgust and Neutral). Accuracy of the developed modelis presented using confusion matrix. The Recognition outcome of the combined MFCC and LPCC extractedfeatures are compared with the isolated results. The maximum accuracy rate that reaches by using the combination of feature extraction method is 88.59% for non linear RBF (Radial Basis Function) kernel SVM.
Authors and Affiliations
Soma Bera , Shanthi Therese , Madhuri Gedam
Keywords
Related Articles
- EP ID EP127463
- DOI -
- Views 115
- Downloads 0
Jamming Anticipation and Convolution through Immaculate Hiding Process of Packets
Cached data not only replies local access, but also replies data request issued from other nodes. Wireless Mesh Networks (WMNs) have emerged as an important technology in building next generation fixed wire...
A Proposed Framework for Integrating Stack Path Identificationand Encryption Informed by Machine Learning as a SpoofingDefense Mechanism
Abstract: Spoofing attacks have been terrorizing the information world for decades; so many methodologieshave been formulated to attempt the eradication of these attacks. This study elaborates on a proposedframewor...
Effective utilisation of ICT is the key to Business success: Case Study of SMEs in Zimbabwe
The focus of this paper is to analyse the utilisation of information communication technology (ICT) among small and medium enterprises (SMEs) in Zimbabwe, as one of the major factors contributing to their limited g...
PARS: Position based Anonymous Routing Scheme in MANETs
Abstract: Mobile Ad hoc Networks (MANET) comprises of self-organizing autonomous mobile nodes. MANETs are open and decentralized. Therefore they are vulnerable to malicious entities. MANETs are multi-hop wireless n...
Regression Test Case Selection &PrioritizationUsing Dependence Graph and Genetic Algorithm
Abstract: Regression testing is very important process in software maintenance. Unfortunately, it iscostly and time consumingto allow for the re-execution of all test cases during regression testing. The challenge...