The Implementation of Speech Recognition Using Kinect Technology

Aulia, Siska and Lifwardaa, - and Veronica, Vera (2017) The Implementation of Speech Recognition Using Kinect Technology. International Conference of Applied Science on Engineering, Business, Linguistics and Information Technology . pp. 375-383. ISSN 2598-2532

[img]
Preview
PDF
737Kb

Abstract

— The implementation of speech recognition processing has grown rapidly so it can be used for various applications such as voice recognition technology using Kinect. Kinect is a motion sensor that has been equipped with speech recognition / voice detection. Kinect has multi-array microphones and produces analog signals. The introduction of speech through kinect technology is a human machine technology that is Automatic Speech Recognition (ASR). The initial stages of the development of speech recognition through kinect is to create a voice detection function and used as a speech recognition database. The speech recognition process uses Mel Frequency Cepstrum Coefficient as a feature extraction, to distinguish one voice from another. Then the Hidden Markov Model (HMM) as a training and pattern recognition. The speech recognition process uses this kinect by searching for the highest probability of a word with a user's spoken word using the HMM method. The result of the test that has been done on 10 words with 25 sources, got the result for speech recognition or trained word of 100%, then the trained source is 82% and the untrained source is 54.4%. From this test proved the system can do word recognition from sound sources that are pronounced with Kinect technology.

Item Type:Article
Subjects:G Technology > GI Electrical engineering. Electronics Nuclear engineering
G Technology > GM Information Technology
Divisions:Luar PNP > Journal, Proceeding
ID Code:2144
Deposited By:Rahmat Hidayat
Deposited On:04 Apr 2018 06:56
Last Modified:23 Dec 2018 08:01

Repository Staff Only: item control page