Abstract
We study a variational model for image amodal completion, i.e., the recovery of missing or damaged portions of a digital image by technics inspired by the well-known amodal completion process in human vision. Representing the image by a real-valued function and following an idea initially proposed in [32], our approach consists in finding a set of interpolating level lines which is optimal with respect to an appropriate criterion. We prove that this method is theoretically well-founded and we show the equivalence with a more classical approach based on a direct interpolation of the function.
| Original language | English |
|---|---|
| Pages (from-to) | 211-252 |
| Number of pages | 42 |
| Journal | Rendiconti del Seminario Matematico dell 'Universita' di Padova/Mathematical Journal of the University of Padova |
| Volume | 116 |
| Publication status | Published - 2006 |
| Externally published | Yes |