Evaluation of the Robustness of Visual Behaviors through Performance Characterization

The robustness of visual behaviors implemented in an active vision system depends both on the vision algorithms and control structure. To improve robustness an evaluation of the performance of the several system algorithms must be done. Performance analysis can be done within the framework of control theory because notions such as stability and controllability can contribute to a better understanding of the algorithms and architectures weaknesses. In this chapter we discuss the generation of reference test target trajectories and we characterize the performance of smooth pursuit and vergence. The responses to motion are used to accurately identify implementation problems and possible optimizations. The system evaluation leads to solutions to enhance the global performance and robustness.