Showing 1–1 of 1 results for author: Kurozumi, T

A quick search method for audio signals based on a piecewise linear representation of feature trajectories

Authors: Akisato Kimura, Kunio Kashino, Takayuki Kurozumi, Hiroshi Murase

Abstract: This paper presents a new method for a quick similarity-based search through long unlabeled audio streams to detect and locate audio clips provided by users. The method involves feature-dimension reduction based on a piecewise linear representation of a sequential feature trajectory extracted from a long audio stream. Two techniques enable us to obtain a piecewise linear representation: the dyna… ▽ More This paper presents a new method for a quick similarity-based search through long unlabeled audio streams to detect and locate audio clips provided by users. The method involves feature-dimension reduction based on a piecewise linear representation of a sequential feature trajectory extracted from a long audio stream. Two techniques enable us to obtain a piecewise linear representation: the dynamic segmentation of feature trajectories and the segment-based Karhunen-Lóeve (KL) transform. The proposed search method guarantees the same search results as the search method without the proposed feature-dimension reduction method in principle. Experiment results indicate significant improvements in search speed. For example the proposed method reduced the total search time to approximately 1/12 that of previous methods and detected queries in approximately 0.3 seconds from a 200-hour audio database. △ Less

Submitted 22 October, 2007; originally announced October 2007.

Comments: 20 pages, to appear in IEEE Transactions on Audio, Speech and Language Processing

Journal ref: IEEE Transactions on Audio, Speech and Language Processing, Vol.16, No.2, pp.396-407, February 2008.