In this paper we describe a corpus annotation project based on crowdsourcing technique that performs orthographic transcription of Kalam'DZ corpus (Bougrine et al., 2017c). This latter is a speech corpus dedicated to Arabic Algerian dialectal varieties. The recourse to crowdsourcing solution is deployed to avoid time and cost consuming solutions that involves experts. Since Arabic dialects have no standard orthographic, we have fixed some guidelines that helps crowd to get more normalized transcriptions. We have performed experiments on a sample of 10% of KALAM'DZ corpus, totaling 8.75 hours. The quality control of the output transcription is ensured within three stages: Pre-qualification of crowd, online filtering and in lab validation and revision. A baseline resource is used to evaluate both first stages. It consists on 5% of the targeted dataset transcribed by well trained transcribers. Our results confirm that the crowdsourcing solution is an effective approach for speech dialect transcription when we deal with under-resourced dialects. Before the validation of the well trained transcribers the accuracy of transcriptions reached 74.38. In addition, we present a set of best practices for crowdsourcing speech corpus transcription.
展开▼
