A new approach for optical flow (image velocity) fieldscomputation is presented using computational neural networks. The computational procedure consists of three stages: estimation of the parameters of the neural ne...
详细信息
A new approach for optical flow (image velocity) fieldscomputation is presented using computational neural networks. The computational procedure consists of three stages: estimation of the parameters of the neural network model, dynamic measurement of the perpendicular velocity components of the contours or region boundaries and computation of the image velocityfields. The parameters are estimated by comparing the energy function of the neural network with a constrained error function. The nonlinear velocity fields computation method is then carried out iteratively by using a dynamic algorithm to minimise the energy function simultaneously with the dynamic measurement of the perpendicular velocity components by a dynamic procedure. Experiments generate velocityfields that are meaningful and consistent with visual perception.
暂无评论