Curvature Regularization for Resolution-Independent Images

A resolution-independent image models the true intensity function underlying a standard image of discrete pixels. Previous work on resolution-independent images demonstrated their efficacy, primarily by employing regularizers that penalize discontinuity. This paper extends the approach by permitting the curvature of resolution-independent images to be regularized. The main theoretical contribution is a generalization of the well-known elastica energy for regularizing curvature. Experiments demonstrate that (i) incorporating curvature improves the quality of resolution-independent images, and (ii) the resulting images compare favorably with another state-of-the-art curvature regularization technique.