Unified approach to reconstruction and modification of motion and image data

This article describes an approach based on hierarchical radial basis functions for the reconstruction and modification of motion/image data, which are formulated as problems of scattered data interpolation. Assuming little about the data and choosing highly scalable/well adjustable basis functions, the approach can handle both the reconstruction and the manipulation in a unified and widely applicable manner: standard approaches are often applicable, but have limited scope due to their constrained assumptions about the number or location of the data points. Reconstruction of captured motion data and warping of face images are demonstrated.