CENSREC-4: An evaluation framework for distant-talking speech recognition in reverberant environments

Acoustical Science and Technology

Norihide Kitaoka²  Yuki Denda^1,12  Satoru Tsuge⁹  Takanobu Nishiura⁵  Masakiyo Fujimoto⁴  Takahiro Fukumori⁵  Satoshi Nakamura^1,13  Tetsuya Takiguchi¹  Masato Nakayama^1,10  Takeshi Yamada^1,11  Chiyomi Miyajima²  Shigeki Matsuda^1,13  Kazuya Takeda²  Shingo Kuroiwa³  Kazumasa Yamamoto⁶  Satoshi Tamura⁷  Tetsuji Ogawa⁸ 

[1] Kobe University;Nagoya University;Chiba University and National Institute of Information and Communications Technology;NTT Communication Science Laboratories, NTT Corporation, “Keihanna Science City,”;Ritsumeikan University;Toyohashi University of Technology;Gifu University;Waseda University;Daido University;Kinki University;University of Tsukuba;Murata Machinery, Ltd.;National Institute of Information and Communications Technology, “Keihanna Science City,”

关键词: Reverberant speech database;    Reverberant speech recognition;    Various recording environments;    Room impulse response;    Evaluation framework;   

DOI  :  10.1250/ast.32.201

学科分类：声学和超声波

来源: Acoustical Society of Japan

【 摘 要 】

References(18)Cited-By(1)We have been distributing a new collection of databases and evaluation tools called CENSREC-4, which is a framework for evaluating distant-talking speech in reverberant environments. The data contained in CENSREC-4 are connected digit utterances as in CENSREC-1. Two subsets are included in the data: “basic data sets” and “extra data sets.” The basic data sets are used for evaluating the room impulse response-convolved speech data to simulate the various reverberations. The extra data sets consist of simulated data and corresponding real recorded data. Evaluation tools are presently only provided for the basic data sets and will be delivered to the extra data sets in the future. The task of CENSREC-4 with a basic data set appears simple; however, the results of experiments prove that CENSREC-4 provides a challenging reverberation speech-recognition task, in the sense that a traditional technique to improve recognition and a widely used criterion to represent the difficulty of recognition deliver poor performance. Within this context, this common framework can be an important step toward the future evolution of reverberant speech-recognition methodologies.

【 授权许可】

Unknown