Abstract :
With the increase of crime and terrorism, the use of video surveillance for security purposes has attracted a great deal of attention. It is crucial there to realize flexible and intelligent vision systems, and it is also essential for developing a variety of practical intelligent systems for inspection, human computer interaction, robotics, and behavior mining. The present paper addresses the theoretical research and applications developed thus far in working toward this goal. First, a theory of feature extraction is briefly introduced in a general framework of pattern recognition. Next, a scheme of adaptive (trainable with learning capability) vision system is presented, which comprises two stages of feature extraction, namely, Higher-order Local Auto-Correlation (HLAC) or its extension CHLAC (Cubic HLAC) and multivariate data analysis. Several interesting applications are demonstrated, showing the effective performance.
Keywords :
computer vision; feature extraction; national security; pattern recognition; terrorism; video surveillance; adaptive vision systems; behavior mining; crime; feature extraction; flexible vision systems; higher-order local auto-correlation; human computer interaction; inspection; intelligent vision systems; multivariate data analysis; pattern recognition; robotics; security purposes; terrorism; video surveillance; Feature extraction; Human computer interaction; Human robot interaction; Inspection; Intelligent robots; Intelligent systems; Machine vision; Robot vision systems; Terrorism; Video surveillance;
