Fast text searching: allowing errors
Communications of the ACM
Query by humming: musical information retrieval in an audio database
Proceedings of the third ACM international conference on Multimedia
Towards the digital music library: tune retrieval from acoustic input
Proceedings of the first ACM international conference on Digital libraries
Melodic matching techniques for large music databases
MULTIMEDIA '99 Proceedings of the seventh ACM international conference on Multimedia (Part 1)
Evaluation of a simple and effective music information retrieval method
SIGIR '00 Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval
A practical query-by-humming system for a large music database
MULTIMEDIA '00 Proceedings of the eighth ACM international conference on Multimedia
Looking for new, not known music only: music retrieval by melody style
Proceedings of the 4th ACM/IEEE-CS joint conference on Digital libraries
QueST: querying music databases by acoustic and textual features
Proceedings of the 15th international conference on Multimedia
Compacting music signatures for efficient music retrieval
EDBT '08 Proceedings of the 11th international conference on Extending database technology: Advances in database technology
An efficient frequent melody indexing method to improve the performance of query-by-humming systems
Journal of Information Science
Song search and retrieval by tapping
AAAI'05 Proceedings of the 20th national conference on Artificial intelligence - Volume 4
Online music search by tapping
Ambient Intelligence in Everyday Life
Hi-index | 0.00 |
Effective use of multimedia collections requires efficient and intuitive methods of searching and browsing. This work considers databases which store music and explores how these may best be searched by providing input queries in some musical form. For the average person, humming several notes of the desired melody is the most straightforward method for providing this input, but such input is very likely to contain several errors. Previously proposed implementations of so-called query-by-humming systems are effective only when the number of input errors is small. We conducted experiments which revealed that the expected error rate for user queries is much higher than existing algorithms can tolerate. We then developed algorithms based on approximate matching techniques which deliver much improved results when comparing error-filled vocal user queries against a music collection.