START Conference Manager    

SCAnnAL – An Automatic Speech Corpus Annotator for African Speech Corpora

Moses Ekpenyong, Eno-Abasi Urua and Aniefon Akpan


Categories

category:  Poster
Session:  6 December Session P4: African Languages Poster Session

Additional Fields

 
Abstract:   Today, thousands of annotated speech corpora exist worldwide and demand for richly annotated corpora is fast growing, but the process accompanying the segmentation and labeling of corpora has slowed research progress for African languages due to the limitations of current annotation Toolkits to satisfy the challenges African speech systems present. We introduce SCAnnAL, a Toolkit for automatic speech annotation that automates the annotation process by accepting raw audio files, segments the waveforms and finally dumps labels into created segments. SCAnnAL is currently being refined for accuracy and is certain to put an end to the intractable procedure of speech annotation.

 
Resume:   Mfịn ami, mme tọsịn adianañkpadia ikọ eba ke ekondo, ñko enekke eyem mme adianañkpadia ikọ ntom ke ise ikọ eti eti. Daña esaña esep enyVñ ewet anyịñ, anam nduuñọ ke mme usem Afrika anyọnyọọñ sia se ekama enam utom ado anana akeene ñkpọ ñnyan ubọk aabañake mme usem Afrika. Imiben SCAnnAL, anamidem akebe ñkpọ usep ñnyVñ ñwet anyịñ ikọ ndọñ ke ise ikọ iwọt. SCAnnAL akeme adidat utatañ ikọ, asep, anyVñ awet anyịñ ke mbaak ikọ. EsVk enanam utom ke SCAnnAL ma akan anọ nneke ibọọrọ anyVñ ayọ mfịna aasañake ke adisep ikọ ndon ke ikpeghe.

File(s)

[Paper (PDF)]  

START Conference Manager (V2.61.0 - Rev. 5964)