Neural modeling of polynomial image interpolation

In this paper, a neural implementation is suggested for polynomial based image interpolation techniques such as bilinear, bicubic and cubic Spline techniques. The performance of the suggested neural image interpolation algorithm is compared to those of traditional polynomial based image interpolation techniques and to warped distance image interpolation techniques. In this implementation, the modeling of the polynomial interpolation algorithm is performed with a fixed number of neurons contrary to the traditional interpolation techniques, which depend on the order of the interpolation basis function. The proposed approach is efficient from the mean square error (MSE) point of view and the computation complexity point of view by reducing the number of computations required. Simulation results show that the suggested implementation has approximately the same peak signal-to-noise ratio (PSNR) as the traditional techniques with a less computational burden due to the fixed size of the neural networks implemented.