RUS  ENG
Full version
JOURNALS // Informatics and Automation // Archive

Tr. SPIIRAN, 2016 Issue 44, Pages 98–113 (Mi trspy857)

This article is cited in 4 papers

Methods of Information Processing and Management

An Analysis of Perspectives for Using High-Speed Cameras in Processing Dynamic Video Information

D. V. Ivankoa, A. A. Karpovb

a ITMO University (Saint Petersburg National Research University of Information Technologies, Mechanics and Optics)
b St. Petersburg Institute for Informatics and Automation of the Russian Academy of Sciences (SPIIRAS)

Abstract: In this paper, we review the actual and perspective areas of use of high-speed video cameras. We discuss the possibility of applying high-speed cameras in the field of human-computer interaction to detect dynamic video information (including visual speech). We also describe main tasks, which can be solved with high-speed cameras, such as: automatic lip-reading, eye blink detection, facial micro-expression recognition, etc. We identify potential challenges associated with the introduction of high-speed video cameras and analyze the conditions of research area. Besides, we analyze state-of-the-art in the field at the moment and prove that there is an urgent need for further scientific and technical developments in this area. We propose some advanced applications and tasks in the human-computer interaction domain, where high-speed video capturing can be useful, such as audio-visual continuous speech recognition and automatic reading speech by lips. In further research, we will implement such a multimodal system for audio-visual Russian speech recognition using a microphone and a high-speed video camera JAI Pulnix.

Keywords: high-speed video camera; computer vision; audio-visual speech recognition; audio-visual data corpus; lip-reading; dynamic video information.

UDC: 004.5

DOI: 10.15622/sp.44.7



Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024