National and Kapodistrian University of Athens
Study and application of acoustic information for the detection of harmful content and fusion with visual information
Abstract
dc:descriptionThis thesis aims at investigating and developing techniques for content based segmentation and classification of multimedia files based on audio information. An introductory investigation of several audio features is fist presented into the context of their respective classification performance. Then a novel method for speech music discrimination of radio broadcasts is proposed. In the sequel an algorithm for locating the parts of an audio stream that contain music (i.e. music tracking) is presented. Another major part of this thesis is related to the development of a computationally efficient audio segmentation algorithm. In the sequel a multi class classification scheme for audio segments from films is proposed in this thesis. The audio classes were selected to describe both violent and non violent content. In the final part the focus on this thesis shifts on the task of recognizing the emotional state of the speaker given a movie speech segment. Finally, an emphasis in this thesis was to develop annotated databases which were used both for training and evaluating the several segmentation and classification methods. To this end, two different types of databases have been used. One from radio recordings which was used in the speech music discrimination task and another one from a number of movies.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Giannakopoulos, Theodoros
- Γιαννακόπουλος, Θεόδωρος
Subjects
dc:subject × 14- Ταξινόμηση ήχων
- Ηχητική κατάτμηση
- Ανάλυση πολυμεσικής πληροφορίας
- Εντοπισμός βίας
- Αναγνώριση προτύπων
- Audio classification
- Audio segmentation
- Multimedia analysis
- Violence detection
- Pattern recognition
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/23952