Multi-view Shape and Texture Learning for Stereo Finite-Element Digital Image Correlation
Résumé
In order to make use of a Stereo Finite-Element Digital Image Correlation framework, one has first to perform a calibration procedure that encompasses the calibration of the cameras and of the specimen shape. This last step is not always straightforward and often requires some kind of regularisation. In the current work, we propose to measure the shape and the texture attached to the specimen and show that it allows to drastically reduce the ill-posedness of the shape measurement problem.