Adaptive Neighborhood Interpolation of Noisy Images on Quincunx Grid

In this paper we present a novel spatially adaptive interpolation technique for grey-scale images corrupted by noise. The problem arises as interpolation of data measured by the sensor of digital photo camera. A locally adaptive interpolation is considered to perform on the quincunx grid. The typical example of such a grid refers to the green colour channel of the Bayer colour …lter array. The technique is based on the directional anisotropic denoising which is employed by local polynomial approximation (LPA). The adaptivity to data is provided by the multiple hypothesis testing called the intersection of con…dence intervals (ICI) rule which is applied for adaptive selection of varying scales (window sizes) of LPA. The adaptive interpolation employs spatial information obtained by ICI for denoising estimates. This joint denoising-interpolation scheme provides more e¢ cient utilization of data then in the classical image restoration chain where restoration of an image is divided on separated steps. Simulations show the e¢ ciency of the proposed joint denoising and interpolation technique.