Reliability Evaluation of Audio Signal Recognition

Autor:	Evgeny Gershikov, Shiran Gabler
Rok vydání:	2018
Předmět:	Signal-to-noise ratio Audio signal Noise measurement business.industry Computer science Speech recognition Spectrogram Usability Mel-frequency cepstrum Fingerprint recognition business Reliability (statistics)
Zdroj:	2018 IEEE XXV International Conference on Electronics, Electrical Engineering and Computing (INTERCON).
DOI:	10.1109/intercon.2018.8526407
Popis:	In this paper we propose reliability measures for algorithms that recognize an audio signal within a database of music recordings when only a short segment of it is available. Using these measures, we test the reliability of two algorithms: one based on spectrogram peaks and one based on MFCC features. We compare the performance of the two methods and conclude about the usability and usefulness of our evaluation.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::9a46035a5f6a1e9795564c6ed6c5d0fb https://doi.org/10.1109/intercon.2018.8526407 Zobrazit plný text záznamu