ABSTRACT
Real time implementation of the Wiener filters for noisy TV imagery restoration is given at the pixel level employing a massive parallel neurochip. Furthermore, mimicking the human visual system, we use the Maxican hat Wavelet Transform (WT) to compress the TV imagery and then we can efficiently restore those dominant scales information for channel distortion in satellite and/or fiberoptics communications. We have coined the name Wiener-Hopfield Neurochip for such a class of LMS restoration neurochips which can solve a large class of filtering in the digital domain.
