ABSTRACT
TV noise is distributed on large region display, which likes a mix of complex noises. The noise type can not be defined by a noise model. As the signal-noise ratio becomes weak, the noise level increases on display accordingly. To improve the image quality, the signal processing for noise reduction had been developed in the past literatures. Nevertheless, the noise problem for large screen TV still can not be completely resolved so far. The well-known median filter and its improved methods can provide a good feature to remove impulse noise signals [1]. However, they are not effective to reduce the large region TV noise. Recently, there are many highly efficient techniques [2-4] that discuss a new filtering approach with wide information in order to remove the noise completely. But their filtering performance is not satisfied for users since the screen noises are still existed. For large region TV noises removal, the performance of the temporal processing is generally better than that of spatial processing due to more information used. In this study, a novel adaptive algorithm is presented to improve the filtering performance based on the adaptation of temporal and spatial processing, particularly for large region TV noise.
