Study of automatic captioning based on unified modeling of spontaneous speech recognition and automatic editing

2015 Fiscal Year Final Research Report

• Project/Area Number: 25730112
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing
• Research Institution: Kyoto University
• Principal Investigator: Akita Yuya 京都大学, 経済学研究科, 講師 (90402742)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 音声認識 / 自動整形 / 話し言葉 / 字幕

Outline of Final Research Achievements

A variety of redundant and colloquial expressions are observed in spontaneous speech such as classroom lectures and academic talks. Thus, automatic speech recognition (ASR) systems need to cover these kinds of spontaneous expressions to make accurate transcripts, while these expressions in the transcripts are then edited and/or removed for better captions. In this research, models of ASR and automatic editing are considered as transformation to/from the style of spontaneous speech. Characteristics of spontaneous speech are modeled, then the model is applied to ASR and automatic editing. We have developed an automatic captioning system based on this framework for lectures.

Free Research Field

音声情報処理