Bioinspired temporal filter modeling for motion estimation

This paper models a set of temporal filters biologically plausible and their hardware implementation, being these filters the first processing stage from a robust neuromorphic motion estimation system. After designing and implementing the filters is constructed a confidence measure which modules the estimation density data using the previously implemented architecture. The global system is examined using several stimuli and error metrics showing a good behavior of the system implemented. This system has many applications regarding tracking, pattern recognition, robotics and real-time signal processing.