Interactive retrieval of spoken content optimizing by LDA algorithm

Abstract

Communication between the human are performed by using different languages. Language is either in written or spoken form. In our paper, as an author we are representing a framework that gives output as a description for audio file using signal processing. The output is in the text format which is derived by examining the audio content and providing the description of audio in the text format. So the study of translation of audio into text goes increasing. The framework is distributed into two sections called training and testing section. The training section is train the audio with its description like speech conversation present in that audio. This data is stored in the system using database with features of scenario of audio. Another section is testing section. The testing section test the audio file and retrieve the output as description of audio comparing audios stored into database (i.e. in training section). Using Latent Dirichlet allocation processing sentences are generated from objects and their activities.