A Framework of Multi-modal Corpus for Mandarin Learning

Abstract

The multi-modal corpus is established on the "theory of speech”. In this paper, a framework of multi-modal corpus for mandarin learning system is presented. The system included five layers which are the physical storage layer, the data description layer, the network layer, the filter layer and the user layer. This system framework is based on the J2EE technologies. The principles of data structure design and the information retrieval model are discussed. This multi-modal corpus covered the entire speech activity that included language, sound, image and action, so it is suitable for the computer-assisted language learning.