The majority of recent work in analysis of visual surveillance have focused on automated perception aspects based on computer vision technology. However, little attention has been paid to higher level semantic analysis upon such perception data. In this dissertation, we approach the higher level semantic analysis from the artificial logical reasoning stance. First, we propose the use of logic programming and subjective logic for knowledge and uncertainty representation. Second, we propose inference schemes for contradiction handling, ambiguous rule modeling, bidirectional conditional reasoning and diagnostic abduction. Each of fore-mentioned aspects are presented with case studies to show the feasibility of the proposed reasoning framework in visual surveillance.
«
The majority of recent work in analysis of visual surveillance have focused on automated perception aspects based on computer vision technology. However, little attention has been paid to higher level semantic analysis upon such perception data. In this dissertation, we approach the higher level semantic analysis from the artificial logical reasoning stance. First, we propose the use of logic programming and subjective logic for knowledge and uncertainty representation. Second, we propose infere...
»