Anomaly-based annotation error detection in speech-synthesis corpora

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F17%3A43931740" target="_blank" >RIV/49777513:23520/17:43931740 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1016/j.csl.2017.04.007" target="_blank" >http://dx.doi.org/10.1016/j.csl.2017.04.007</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.csl.2017.04.007" target="_blank" >10.1016/j.csl.2017.04.007</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Anomaly-based annotation error detection in speech-synthesis corpora

Popis výsledku v původním jazyce

We investigate the problem of automatic detection of annotation errors in single-speaker read-speech corpora used for speech synthesis. For the purpose of annotation error detection, we adopt an anomaly detection framework in which correctly annotated words are considered as normal examples on which the detection methods are trained. Misannotated words are then taken as anomalous examples which do not conform to normal patterns of the trained detection models. The results with F1 score being almost 89% show that anomaly detection could help detecting annotation errors in read-speech corpora for speech synthesis. We show that the automatically reduced feature sets achieve statistically similar results as the hand-crafted feature sets. We show that a reasonably good detection performance could be reached with using significantly fewer examples during the detector development phase. We also propose a concept of a voting detector – a combination of anomaly detectors in which each “single” detector “votes” on whether or not a testing word is annotated correctly, and the final decision is then made by aggregating the votes. Our results show that the voting detector has a potential to overcome each of the single anomaly detectors. Furthermore, we compare the proposed anomaly detection framework to a classification-based approach (which, unlike anomaly detection, needs to use anomalous examples during training) and we show that both approaches lead to statistically comparable results when all available anomalous examples are utilized during detector/classifier development. However, when a smaller number of anomalous examples are used, the proposed anomaly detection framework clearly outperforms the classification-based approach. A final listening test showed the effectiveness of the proposed anomaly-based annotation error detection for improving the quality of synthetic speech.

Název v anglickém jazyce

Anomaly-based annotation error detection in speech-synthesis corpora

Popis výsledku anglicky

We investigate the problem of automatic detection of annotation errors in single-speaker read-speech corpora used for speech synthesis. For the purpose of annotation error detection, we adopt an anomaly detection framework in which correctly annotated words are considered as normal examples on which the detection methods are trained. Misannotated words are then taken as anomalous examples which do not conform to normal patterns of the trained detection models. The results with F1 score being almost 89% show that anomaly detection could help detecting annotation errors in read-speech corpora for speech synthesis. We show that the automatically reduced feature sets achieve statistically similar results as the hand-crafted feature sets. We show that a reasonably good detection performance could be reached with using significantly fewer examples during the detector development phase. We also propose a concept of a voting detector – a combination of anomaly detectors in which each “single” detector “votes” on whether or not a testing word is annotated correctly, and the final decision is then made by aggregating the votes. Our results show that the voting detector has a potential to overcome each of the single anomaly detectors. Furthermore, we compare the proposed anomaly detection framework to a classification-based approach (which, unlike anomaly detection, needs to use anomalous examples during training) and we show that both approaches lead to statistically comparable results when all available anomalous examples are utilized during detector/classifier development. However, when a smaller number of anomalous examples are used, the proposed anomaly detection framework clearly outperforms the classification-based approach. A final listening test showed the effectiveness of the proposed anomaly-based annotation error detection for improving the quality of synthetic speech.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

<a href="/cs/project/GA16-04420S" target="_blank" >GA16-04420S: Kombinované využití fonetických a korpusově založených postupů při odstraňování rušivých jevů v řečové syntéze</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2017

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

COMPUTER SPEECH AND LANGUAGE

ISSN

0885-2308

e-ISSN

—

Svazek periodika

46

Číslo periodika v rámci svazku

November

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

35

Strana od-do

1-35

Kód UT WoS článku

000407609600001

EID výsledku v databázi Scopus

2-s2.0-85019202074