Accuracy control of reconstruction errors in volume slicing

Slicing is a most powerful and useful tool for the exploration of volume data. In the core of the slicing algorithm is a resampling process that produces a set of new samples constituting the desired slice. Reconstruction is imperative whenever an image or a volume needs to be resampled as a result of an affine or perspective transformation, texture mapping, or volume slicing. Traditionally, this reconstruction is performed by using trilinear or cubic interpolation. However, the resulting images suffer from significant artifacts and there exists no way to guarantee some error level at every pixel. We employ a new method for the characterization and measurement of reconstruction error that is based on spatial domain error analysis. Our approach supports also position-adaptive and data-adaptive reconstruction which adjust filter size to the location of resampling point and the data in its vicinity respectively