A Platform for Managing Term Dictionaries for Utilizing Distributed Interview Archives

This paper proposes a platform that aims to support the whole process and facilitate archiving tasks at museums and galleries. When we try to preserve tacit knowledge or skills of artists or masters by interview and utilize interview videos, it is important to support the whole process of archiving; capturing interview, digitizing it to a machine readable format, authorizing metadata, and exporting them to be accessed from others. To interoperate distributed archives, rich meta information is required. In this paper, we focus our tools to facilitate authoring metadata of interview, such as transcription and annotation task for interview, with speech processing or natural language processing functions. The important point now is that the performance of the speech processing or text processing for specific domain depends on well-fitted language models to it. We, therefore, design our tools as an networked software package that can download up-to-date term dictionary and language models f...

• Type: Conference
• Year: 2006
• Where: CAISE
• Authors: Kenro Aihara, Atsuhiro Takasu
• Keywords: CAISE 2006, Information Management, Language Models, Speech Processing, Text Processing