BUILDING AND TRAINING RADIOGRAPHIC MODELS FOR FLEXIBLE OBJECT IDENTIFICATION FROM INCOMPLETE DATA

The authors address the problem of identifying the projection of an ob ject from incomplete data extracted from its radiographic image. They assume that the unknown object is a particular sample of a flexible ob ject. Their approach consists first in designing a deformation model a ble to represent and to simulate a great variety of samples of the fle xible object radiographic projection. This modellisation is achieved u sing a training set of complete data. Then, given the incomplete data, the identification task consists in estimating the observed object us ing the deformation model. The proposed modelling extracts from the tr aining set, not only the deformation modes, but also other relevant in formation (such as probability distributions on the deformations, rela tions between deformations) to use it to regularise the identification step.