|
Department of
Computational Perception
|
|
Mögliche Themen für Praktika und
Diplomarbeiten
Vorbemerkung
Die folgenden Themen sind nach Typ (PR =
Projektpraktikum, Praktikum aus Informatik oder Praktikum aus
Pervasive Computing; DA = Diplomarbeit) geordnet. Diese Typen sind
aber nur als grobe Richtschnur gedacht. Manche Praktika sind auch
zur Durchführung in Gruppen geeignet, und Praktikumsthemen können
auch zu Diplomarbeiten ausgebaut werden.
Wir sind außerdem jederzeit für neue Themenvorschläge offen – wenn
Sie an einer Frage interessiert sind, die Sie im Rahmen eines
Praktikums oder einer Diplomarbeit bearbeiten wollen, dann melden
Sie sich!
Beispiele für mögliche weitere Themen:
- diverse interessante Plugins für Media Players und
mp3-Player
- Erkennung verschiedener musikalisch relevanter Muster in Musik
(Melodie, Akkorde, ...)
- usw. usw.
Music Information Retrieval
Feature Extraction, Similarity Measures,
Playlist Generation
|
|
Typ |
Thema |
Betreuer |
|
PR
DA |
Investigating modern interpretations of traditional (medieval)
music |
M. Schedl
B.
Niedermayer |
PR
DA |
Discriminating Speech and Singing Voice
|
K.
Seyerlehner |
| PR |
C++ implementation of factorization-based features |
A. Arzt
B.
Niedermayer |
| PR |
Investigating Skew- and Log-Normal-Distributions to Model Music
Similarity |
T. Pohle |
| PR |
Implementing various context-based music similarity
approaches |
M. Schedl
P. Knees |
|
| PR |
Automatic Tagging and Organizing Music Collections using Audio
Fingerprinting Technology |
M. Schedl |
|
Intelligent User Interfaces and
Information Visualization
Real-Time Music Understanding
|
|
Typ |
Thema |
Betreuer |
|
|
| PR |
The On-Line Performance Worm: Visualising Expressive Music
Performance in Real Time via our Automatic Audio
Tracker |
A. Arzt
G. Widmer |
| DA |
Real-time Identification of Musical Motives in Audio Streams. |
A. Arzt
G. Widmer |
PR
DA |
The Vocal Joystick: Controlling Video Games by "Singing" |
A. Arzt
G. Widmer |
| PR |
Stream-based Score-Performance Alignment |
S.
Flossmann |
PR
DA |
An Interactive Tool for Reconstructing Music Scores from MIDI
Files |
G. Widmer |
| PR |
Prediction of Sustain Pedal Use in Expressive Piano
Performances by Machine Learning |
S.
Flossmann |
| PR |
Estimating single-pitch velocities from harmonic mixtures |
B.
Niedermayer |
| PR |
Adding Computational Intelligence to Disc-JoQey: A Semi-Automatic Application to Segment
and Tag Recordings from Vinyls |
P. Knees |
|
Music Context and User Context
Understanding, Web and Social Mining, Personalization
|
|
Typ |
Thema |
Betreuer |
|
| DA |
User Profiling, Models for user context-based music similarity
estimation |
M.
Schedl
|
PR
DA |
Popular Artist/Song Detection |
M. Schedl |
| PR |
Information Aggregation (e.g. from RSS-feeds, Blogs, Social
Networks, ...) to build a Personalized Music News-System |
M. Schedl |
PR
DA |
Assessing the power of YouTube for music/multimedia information retrieval
(tags, user discussions, playlists, image previews of related
videos, recommendations, ...) |
M. Schedl |
PR
DA |
Extracting and comparing Web-based/collaborative
filtering-based artist similarities across different (Web 2.0)
platforms/services (e.g., last.fm, allmusic.com, myspace, Amazon,
Yahoo! Music, ...) |
M. Schedl |
PR
DA |
Extracting and Visualizing Relations from Music-related Web
pages (band members, albums, songs, ...) |
M. Schedl |
| PR |
Adaptive query reformulation strategies for Web-based MIR |
M. Schedl
P. Knees |
PR
DA |
Novelty
Detection for MIR via Web Information Extraction
(bands, album releases, ...) |
M. Schedl |
PR
DA |
An Efficient Focused
Crawler in Java for Music-related, Large Scale Web
Crawling |
M. Schedl
P. Knees |
PR
DA |
Analyzing the Potential of Web
Page Segmentation to Improve Text-based Artist Similarity
Measures (in Java) |
P. Knees |
PR
DA |
Image Retrieval for Music-related Images (Band Pictures, Album
Covers, Logos, ...) |
M. Schedl |
| PR |
Finding correlations of image properties and
artists/genres |
M. Schedl |
| PR |
Evaluating various Web-based music information systems (w.r.t.
functionality, usability, interfaces, ...) |
M. Schedl |
|
|
Web Mining (General)
|
Typ |
Thema |
Betreuer |
|
|
PR
DA |
Automatic analysis of bookmarks to create a user model
for Web page
recommendation |
M. Schedl |
| PR |
Automatische Erzeugung
von Literaturlisten (BibTeX) aus PDFs (per Web Content
Mining) |
B.
Niedermayer
M. Schedl |
| PR |
Implementing Visual Web Page
Segmentation in Java |
P. Knees |
|
Bild- und Sprachverarbeitung, Biometrie
Miscellaneous
Fragen etc. an
Für inhaltliche Fragen zu den Themen, kontaktieren Sie bitte
direkt den/die entsprechenden Betreuer. Zu allen anderen Fragen
(organisatorische, etc.), kontaktieren Sie bitte eine der folgenden
Personen.
Gerhard
Widmer
Tel. 2468-1510
Josef Scharinger
Tel. 2468-8898
Markus Schedl
Tel. 2468-1512
last edited by ms at
2010-03-11