<title>Intermediate image interpolation using polyphase weighted median filters</title>

Abstract

A new algorithm for the interpolation of temporal intermediate images using polyphase weighted median filters is proposed in this paper. To achieve a good interpolation quality not only in still but also in moving areas of the image, vectorbased interpolation techniques have to be used. However, motion estimation on natural image scenes always suffers from errors in the estimated motion vector field. Therefore it is of great importance, that the interpolation algorithm possesses a sufficient robustness against vector errors.Depending on the input and output frame repetition rate different cyclically repeated interpolation phases can be distinguished. The new interpolation algorithm uses dedicated weighted median filters for each interpolation phase (polyphase weighted median filters) which are (due to their shift property) able to achieve a correct positioning of moving edges in the interpolated image, even if the estimated vector differs from the true motion vector up to a certain degree. A new design method for these dedicated error tolerant weighted median filters is presented in the paper. Other aspects like e.g. the preservation of fine image details can also be regarded in the design process. The results of the new algorithm are compared to other existing interpolation algorithms.

Combining vector ordering and spatial information for color image interpolation

“…To solve this problem, some median or median filtering based interpolation methods [1,7,8] have been developed. However, in these methods only the relationship between image pixels and no spatial information are considered.…”

“…Although many interpolation algorithms have been developed [1][2][3][4][5][6][7][8], most of the conventional techniques, such as the nearestneighbor interpolation, bilinear interpolation, bicubic interpolation, and the spline-based methods [2][3][4][5][6], only utilize the spatial information of the neighborhood pixels around the interpolation positions and do not take account of the inherent relationship between the neighborhood pixels. To solve this problem, some median or median filtering based interpolation methods [1,7,8] have been developed.…”

Super-Resplution Procedures in Images and Video Sequences Based on Wavelet Atomic Functions

“…So, it is assumed that the images or frames are treated as LR ones, where the promising methods of super resolution to the entire image or area of interest should be employed recovering the data lost during acquisition stage. Finally, reconstructed data present more information for better visual understanding of selected areas, permitting a deeper analysis for various purposes (Baboulaz et al, 2009) There are exist a lot of the algorithms in the SR (Franzen et al, 2001;Chaudhuri et al, 2001), among them, the nearest neighbour methods that employ the interpolation procedure with the closest pixels to approximating point; bi-linear interpolation (Hou et al, 1978) that applies the mean averaging filter for neighbouring pixels in each a central pixel, and in this way obtaining the lost pixels; the bi-cubic algorithm that uses the cubic polynomial function for additional pixels; the methods based on spline technique (Lehmann et al, 1999;Phu et al, 2004) that deform the edges and wave them. Simple interpolation-based methods, such as bilinear or bicubic interpolation, etc.…”

Algorithm of super-resolution in images and video sequences applying wavelets based on atomic functions

“…A lot of the interpolation algorithms have been developed [1,2] in super-resolution (SR), among them are: the nearest neighbor method that uses the interpolation procedure where the pixel, which is the closest to given approximating point is selected; bilinear interpolation [3] that applies the mean averaging filter for pixels along horizontal and vertical axes for each a central pixel; the bi-cubic algorithm [3,4] that employs cubic polynomial function forming additional pixels; the methods based on spline technique [5,6] that deform the edges and wave them. All these methods only apply the spatial pixels information, so, some other methods based on the median filtration or median interpolation have been developed and studied [1,2,6] employing vicinity relations. There exist other algorithms, in particular warping, which is based on re-sampled operation employing the point spread function approximation in rectangular form.…”