Volume & Issue no: Volume 6, Issue 4, July - August 2017
____________________________________________________________________________________________________
Title: |
Speech Emotion Detection Based On Optimistic – DNN (Deep Neural Network) Approach |
Author Name: |
Inderjeet Kaur, Dr. Rakesh Kumar & Palwinder Kaur |
Abstract: |
Abstract
Inrecent years, artificial intelligence like HCI (Human
Computer Interface) has become very hot issue. The additional
speech emotion recognition capabilities, speech based interfaces
could be created more human-centric. The natural languages
do not share the same features and variation in the process of
speech, sound of speech; the recognition of speech accuracy
gets affected with respect to the consumer’s language. The main
aim of this research is to study the design of intonation and
stress for voice emotion in English and to study the influence of
person on emotion of speech’s recognition and accuracy.
Speech information corresponds to the several type of emotions
as Happy, Sad, Surprise, Joy and Aggressive. It was collected
from 32 samples which contain various audio files. The
research paper implements a combined system for detection of
speech emotion by feature extraction using Gamma Tone
Cepstral and Coefficient (GTCC); Enhanced Genetic Algorithm
(EGA) used to extract features based on pitch, energy, filtered
data and frequency. Classification is used to detect emotion in
by supervised approach which have K-mean clustering
algorithm and DNN algorithms. The consequences of the study
defined are reduced features set gave better performance results
as compared to SVM and Hybrid SVM approach.
Keywords: Speech emotion recognition, Feature
extraction, Gamma tone cepstralcoefficient(GTCC),
Enhanced genetic algorithm, K-mean clustering and DNN
algorithm. |
Cite this article: |
Inderjeet Kaur, Dr. Rakesh Kumar & Palwinder Kaur , "
Speech Emotion Detection Based On Optimistic – DNN (Deep Neural Network) Approach" , International Journal of Emerging Trends & Technology in Computer Science (IJETTCS) ,
Volume 6, Issue 4, July - August 2017 , pp.
150-156 , ISSN 2278-6856.
|
| Full Text [PDF] Back to Current Issue | |
NOTE: Authors note that paper cannot be withdrawn at any condition once it is accepted. The Team of IJETTCS advise you, do not submit same article to the multiple journals simultaneously. This may create a problem for you. Please wait for review report which will take maximum 01 to 02 week.