Etienne Le Quentrec
PhD Student
|
PhD supervision
- PhD director
- Mohamed Tajine
- PhD supervisors
Thesis subject :
When an object is numerically captured, the resulting image is pixelated and is described by integral coordinates. A lot of information is then lost. The goal of my thesis is to estimate the geometric features of the initial object (such as its perimeter) knowing only its digitization. More precisely, by modeling the digitization process, we are aiming to guarantee bounds on the estimation error of the geometric features. My work consists in introducing a new family of shapes (including regular shapes and polygons) whose boundary curvature is bounded, in proving the preservation of the topology of the digitization and in guaranteeing the robustness of length estimation and line integral on this family of shape. keyword: digital geometry, topology preservation, length estimation, line integral estimation, curvature.
Publications :