Wavelet-based Audio Fingerprinting Algorithm Robust to Linear Speed Change

摘要：

In this paper, a novel audio fingerprinting scheme for content-based audio retrieval was presented. To efficiently utilize the time-frequency characteristics of audio signal, the 8-levels wavelet transform is adopted to obtain the detail timefrequency features. Experiments show that the proposed scheme not only has good robustness to content-preserving operations and additive white Gaussian noise (AWGN) but also shows highly robustness against large linear speed change. Furthermore, compared to other existing schemes, it requires less storage space and computational costs.

关键词： audio fingerprinting wavelets

作者: Jixin Liu Tingxian Zhang

作者单位: Department of Computer and information science Fujian University of Technology FuZhou, China Wireless Center China TeleCom GuangDong Branch GuangZhou, China

会议类型: 国际会议

会议名称: 2010 International Conference on Bio-inspried System and Signal Processing(2010 IEEE生物系统与信号处理国际会议 ICBSSP 2010)

会议地点: 厦门

会议语种:英文

页码: 73-76

在线出版日期: 2010-10-26（万方平台首次上网日期，不代表论文的发表时间）