Feature Extraction Approach to Enhance Information Retrieval for Musical Media

Dissertant

al-Jafar, Ahmad Dalli Mahir

Thesis advisor

al-Azami, Muayyad A. Fadhil

Comitee Members

al-Umari, Faruq Ahmad
Tayi, Muhammad Mustafa

University

Philadelphia University

Faculty

Faculty of Information Technology

Department

Department of Computer Science

University Country

Jordan

Degree

Master

Degree Date

2013

English Abstract

-In this research, the implementation of advanced music information retrieval system using the MATLAB software program is proposed and analyzed.

This implemented system retrieves musical data based on initially pre-processing the musical data using three main techniques; Discrete Wavelet Transform (DWT), Linear Predictive Coding (LPC) and filtering function, after that, extracting seven features; energy in DWT of the high band, Energy in DWT of the low band, total energy of both bands, zero crossing rate of the high band, zero crossing rate of the low band, autocorrelation of the high band and autocorrelation of the low band from these data and then matching the extracted features with those of training musical data that are stored in a predefined database.

The matching process depends on computing the Minkowski distance among musical data.

Results demonstrate that the implemented music retrieval system can effectively and accurately retrieve the correct musical data based on the chosen testing data and comparing them with predefined ones

Main Subjects

Arts
Information Technology and Computer Science

Topics

• Data processing
• Music

No. of Pages

66

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-544056