Author:
Song Yue,Tao Sha,Ren Yanzhao,Liu Xinliang,Gao Wanlin
Abstract
Abstract
With the popularization of Internet and multimedia technology, a large amount of agriculture-related audio information is rapidly accumulated and widely used. The method of keyword matching retrieval for text data content describing audio files can no longer meet the need of agricultural audio files retrieval in terms of retrieval efficiency and content depth. In view of the above problems in agricultural audio retrieval technology, this paper proposes an audio retrieval method based on inverted index of silence word. In this method, a new hybrid retrieval model combining representation retrieval and semantic retrieval is designed and used. The experimental results show that the recall rate and precision rate of this method are above 99.5%, and the average retrieval time is less than 1s. Therefore, this method greatly improves the efficiency and the accuracy of the audio information retrieval, and can be used for rapid retrieval of agricultural audio.
Reference14 articles.
1. Direct access in content-based audio information retrieval: a state of the art and challenges[C];Heryanto
2. An Inverted Index Based Audio Retrieval Method [J];Zhang;Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology,2013
3. Time-frequency matrix feature extraction and classification of environmental audio signals[J];Ghoraani;IEEE Transactions on Audio, Speech, and Language Processing,2011
4. Music classification via the bag-of-features approach [J];Zhou-yu;Pattern Recognition Letters,2011