Sensors & Transducers
Vol. 242, Issue 3, March 2020, pp. 1-5
_______________
Novel Speech Endpoint Detection Algorithm for Voice Detectors in Interaction of Intelligent Terminals
1, * Jixiang LU and 2 Xiao HAN
1
School of Mechanical Engineering, Northwestern Polytechnical University, Xi'an 710072, China
2 Hamburg University of Technology, Hamburg, 21107, Germany
1 Tel.: +86 13319288957 (CHN)
E-mail: lujixiang@nwpu.edu.cn
Received: 29 February 2020 /Accepted: 25 March 2020 /Published: 31 March 2020
Abstract:
The human-computer interaction technology of intelligent terminals has totally altered from computer-centered to natural-oriented human-centered paradigm. As the most natural interaction mode of human beings, speech interaction has become a hot topic in both research and application fields in recent years. In the practical application environment of speech interaction technology, a signal recognition and processing method is required to get clean speech from the noisy speech. This paper presents a speech enhancement technique which is based on improved spectrum subtraction by using auditory masking effect, and introduces a novel speech endpoint detection algorithm, which overcomes the influence of music noise and extracts pure speech.
Keywords:
Intelligent terminals, Human-computer interaction, Voice detector, Cloud computing.
.__________________________________________________________________________________________