Frédéric Sur

On Noise Prediction in Maps Obtained With Global DIC

A predictive formula giving the measurement resolution in displacement maps obtained using Digital Image Correlation was proposed some years ago in the literature. The objective of this paper is to revisit this formula and to propose a more general one which takes into account the influence of subpixel interpolation for the displacement. Moreover, a noiseless DIC tangent operator is defined to also minimizes noise propagation from images to displacement maps. Simulated data enable us to assess the improvement brought about by this approach. The experimental validation is then carried out by assessing the noise in displacement maps deduced from a stack of images corrupted by noise. It is shown that specific image pre-processing tools are required to correctly predict the displacement resolution. This image pre-processing step is necessary to correctly account for the fact that noise in images is signal-dependent, and to get rid of parasitic micro-movements between camera and specimen that were experimentally observed and which corrupt noise estimation. Obtained results are analyzed and discussed.

Shape Recognition Based on an a Contrario Methodology

The Achilles’ heel of most shape recognition systems is the decision stage, whose goal is to clearly answer the question of whether two shapes look alike or not. In this chapter we propose a method to address this issue, that consists in pairing two shapes as soon as their proximity is unlikely to be observed “by chance.” This is achieved by bounding the number of false matches between a query shape and shapes from the database. The same statistical principle is used to extract relevant shape elements from images, yielding a complete procedure to decide whether or not two images share some common shapes.