Matti Karlsson

Color video signal processing with median filters

Methods of applying median filtering for three-dimensional video signals were studied and are compared. Median operation and RGB- versus YUV-domain processing are discussed. An introduction to various ordering methods in median-based multidimensional filtering is given. The various 3-D median processing methods and the effect of the conversion from the RGB to the YUV domain were tested. The signal processing methods used for testing were noise reduction and interlaced-to-progressive conversion. The test results are given. >

Evaluation of scanning rate up conversion algorithms; subjective testing of interlaced to progressive conversion

The suitability of different scanning formats and upconversion algorithms for improved definition television (IDTV) applications is evaluated. The properties of different scanning formats are discussed. State-of-the-art interlaced-to-progressive conversion algorithms are introduced. Algorithms were evaluated with subjective testing methods, and the results are compared with objective measures. >

Line Rate Upconversion In IDTV Applications

A median-based motion adaptive algorithm has been developed for the interpolation of 625 extra lines between 625 original lines in a normal-definition PAL receiver. Furthermore, a motion detection algorithm suitable for cost-efficient implementation in IDTV (improved-definition television) receivers has been developed. A motion detector is required to determine whether the pixel to be interpolated is moving. It was concluded that the motion adaptive weighted median filtering offers better picture quality than other methods studied, while allowing cost-efficient production. >