Inter-slice Resolution Improvement of Lung 4D-CT via Adaptively Patch Partition and Sparse Representation

Lung four-dimensional computer tomography (4D-CT) data resolution enhancement is helpful for lung cancer accurate radiotherapy. Sparse representation based algorithm has been proposed to reconstruct the resolution enhancement image, and achieved state-of-art performance. In this paper, based on the sparse representation algorithm, we present an adaptively patch partition approach to divide the slices into adaptively scaled patches. This approach will catch more anatomical nuances and improve the reconstruction. The quad tree-based algorithm is employed in our method to partition the slices. Moreover, a jointly intensity-feature homogeneity is defined to determinate the patch division criterion. The effectiveness of the proposed method is demonstrated by the experiments.