Abstract:
Please note: This talk is NOT my qual talk. I will be presenting more detail with a much wider scope. My qual talk will be a subset of this talk with an emphasis on moving object correspondence.

I will begin by briefly outlining the goals of visual surveillance within the scope of the CyberScout-ATV project. I will then propose a breakdown of the tasks to achieve the goals along with prior and related work in each of the subtasks.

So far we have focused on the low-level processing demands of visual surveillance. Specifically, motion detection, classification and moving object correspondence. I will present the details and results of the motion detection and moving object correspondence algorithms.

The current efforts of the CyberScout-ATV surveillance group are focused toward defining and developing high-level processing algorithms. I will outline the directions and present some early results.