One dimensional multi-velocity motion detection neural networks based on lateral inhibitory mechanism

Motion perception is one of the most important components of biological vision system. In this paper, by simulating the lateral inhibitory mechanism of the neurons in the motion area of visual cortex, a new Multiplicative Inhibitory Velocity Detector (MIVD) was proposed. The optimum velocity of this detector is determined by its spatial-temporal joint parameter K. Based on the response disparity of MIVD, two multi velocity fusion Neural Networks (a simple one and an active one) were constructed to detect the velocity of one dimensional motion. The experiments show that the active MIVD Neural Networks with feedback have better detecting result.