IMAGE INTERPOLATION-BASED ON MEDIAN-TYPE FILTERS

Image interpolation based on median-type filters is investigated. We s tudy this problem for two decimation schemes that are most commonly us ed in practice: (1) quincunx lattice and (2) rectangular lattice. For each decimation lattice, we propose a number of interpolation structur es that are based on some median or finite impulse response (FIR)-medi an hybrid filters. Extensive simulations show that the proposed median -based interpolators outperform significantly as compared with linear schemes. To further verify the effectiveness of these median-based int erpolators, we compare them with multilayer neural networks designed t hrough a training process, and still show a preference for the median- based interpolators. As an application, they are used in an interpolat ive Joint Photographic Experts Group (JPEG) coding scheme to suppress blocking artifacts at very low bit rates. (C) 1998 Society of Photo-Op tical Instrumentation Engineers.