Hindi Audio Indexing and Retrieval

The use of internet for unlimited data storage and retrieval has become an essential in day to day life. The retrieval of data is possible with few keywords from the unlimited number of stored data. Since considerable part of stored data are audio files, it is necessary to retrieve the speech from audio files for indexing and retrieval. The proposed work considers converting the audio file to text data for indexing. All over India, one of the most used regional language Hindi audio files is considered for processing. Speech recognition is mandatory in many application areas. The proposed work is carried out as two phases. The first phase is speech recognition of input audio files using Google Speech API and converted to text data. The second phase is retrieving the audio files with specific keywords. The experiment is conducted with input from youtube videos for better results.