Artifact reduction in sinc interpolation using adaptive filtering

We propose a new method for artifact reduction of upsampled multidimensional signals. These artifacts are evident near edges and they are due to the spectral narrowing associated with upsampling. The method is based on first upsampling the signal with conventional optimal sinc interpolation and then applying a local filter that reduces the high frequencies associated with the ringing artifacts along the edges while leaving unchanged the directions orthogonal to them. The method is specially suitable when dealing with medical images that contain small structures, such as thin bones in CT