Enhancement of Speech Signal Using Improved FA-ANFIS Classifier for Hearing AIDS Application

1Dr. N. Shanmugapriya


Research is undergoing in hearing aids application in the sense of dealing with background noise in speech. It aims to understand the speech in the availability of background noise. It acts as challenging environment to enhance the speech understanding during presence of noise. Thus hearing aid applications had introduced to understand the speech even in presence of noise in various environments for end users. It is one type of methodology adopted for improving the user to understand the information in signal. This speech enhancement application helps to improve availability of data. Additionally, it increases the speech signal intelligibility and quality of application. This review introduced the mechanism for improving these above mentioned characteristics in applications such as hearing aids. This could be followed out using enhancement methods namely, Fast Independent Component Analysis algorithm shortly Fast ICA. This helps to reduce the noise presence in speech related signals. This process also involves IDWT techniques where IDWT stands for Improved Discrete Wavelet Transform to do feature selection. Features have been extracted from speech signal that are denoised. AANOVA stands for Advanced Analysis of Variance helps to identify the important feature from the features that are already selected in previous process. Finally, the features which are extracted as important have been given for optimization algorithm in order to get the features in optimized manner. Now lastly collected features are then classifies efficiently using ANFIS. ANFIS stands for Adaptive Neural Fuzzy Interference System is a classification approach. It also called as Improved FA-ANFIS Method. The result obtained in proposed approach proves that it would gives out promisingly better speech intelligibility and the good quality over speech signal in hearing aids commodities as compared to applications already exist. Keywords--- Sp


Speech Signal Enhancement, Fast Independent Component Analysis (Fast ICA), Improved Discrete Wavelet Transform (IDWT), Firefly Algorithm, Improved FA-ANFIS.

Paper Details
IssueIssue 3