Non-linear techniques for video de-interlacing

De-interlacing is the ability to estimate one field in a video image using the other field and is a form of interpolation. Such a task is required in many video applications, such as standards conversion. Unfortunately the performance of de-interlacers based on linear filters is unsatisfactory. An alternative is to use non-linear interpolation. Two non-linear techniques are discussed for this problem: Volterra methods and radial basis function networks. Results illustrating the performance of both non-linear methods are presented along with results for linear techniques. The results are judged bearing in mind that a major consideration in the implementation of such nonlinear systems is their computational complexity.