Literature Review on Speech Based Information Retrieval through web for visually impaired

Abstract

Speech-based content retrieval system facilitates the visually impaired to have an efficient access, search and browsing capabilities to the desired content in audio form. In Speech based content retrieval method, Query is given as voice command to the system; the system has to generate the information based on the spoken word. The information has to be retrieved based on the keyword matching. The resultant information should be converted as a speech. Usual Method used in Speech Recognition (SR) is Hidden Markov Model (HMM), Dynamic Time Warping (DTW) [1]. If the generated system is Speaker Dependent, then it can’t analyze the user’s voice (input) for further process. So the system has to be implemented for Speaker Independent. So that it can generate speech signals and its corresponding text [2]. People with visual disability or impaired can make this type of web search. Keywords: Dynamic Time Warping (DTW), Hidden Markov Model (HMM), Hyperlinks and Algorithm, Information Retrieval, Keyword Matching, Rough-fuzzy Method, Speech Recognition (SR), Speech Synthesis.