A real-time lipreading LSI for word recognition

Author

Nakamura, Kazuhiro ; Murakami, Noriaki ; Takagi, Kazuyoshi ; Takagi, Naofumi

Author_Institution

Center for Inf. Media Studies, Nagoya Univ., Japan

fYear

2002

fDate

2002

Firstpage

303

Lastpage

306

Abstract

In the paper, we present a real-time lip-reading LSI for recognizing spoken words from lip movement. The LSI recognizes up to 8 words based on the hidden Markov model (HMM). The LSI accepts the 256×256 8-bit gray-scale images from a camera, and outputs the 3-bit symbol code of words for 43 images (corresponding to 1.53 s). We present a lip-reading algorithm optimized for hardware implementation. We have designed the lip-reading LSI and fabricated a 4.9 mm×4.9 mm chip using 0.35 μm process via VDEC Rohm. The LSI performs real-time recognition at 40 MHz operation.

Keywords

circuit optimisation; hidden Markov models; image recognition; integrated circuit design; integrated circuit testing; large scale integration; speech recognition equipment; VDEC Rohm process; camera gray-scale images; hardware implementation; hidden Markov model; lip movement; optimized lip-reading algorithm; real-time lip-reading LSI; real-time recognition; spoken word recognition; symbol code output; Cameras; Gray-scale; Hardware design languages; Hidden Markov models; Humans; Image edge detection; Image recognition; Large scale integration; Speech recognition; Vector quantization;

fLanguage

English

Publisher

ieee

Conference_Titel

ASIC, 2002. Proceedings. 2002 IEEE Asia-Pacific Conference on

Print_ISBN

0-7803-7363-4

Type

conf

DOI

10.1109/APASIC.2002.1031592

Filename

1031592