Local-Feature Based Adaptive Reconstruction for Volume Slice

The paper describes the local features within the volume data using first and second order derivatives around interpolated points,and then classifies the space into constant zones,linear zones,and non-linear zones according to the types of local feature.Accordingly,different filters are employed to reconstruct accurate volume slice.The experiment results show that the method can generate higher quality images than the traditional linear filter and has less computational load than using pure complex filters.