Titre: Topology-guided Visualization of Scalar Datasets Auteurs: Léo Allemand-Giorgis (orateur), G.-P. Bonneau, S. Hahmann, Université de Grenoble, LJK et INRIA Résumé: Topological approaches for the visualization of complex datasets have recently emerged as a successful area of research. For scalar datasets, topological combinatorial structures such as the Reeb graph, the contour tree, the Morse and the Morse-Smale complexes have found practical applications for defining transfer functions in volume rendering, simplifying meshes or extracting and visualizing the main critical points of a dataset. This talk is related to Morse-Smale complexes which are topological structures connecting critical points in a scalar dataset. Previous works have handled the efficient computation and simplification of Morse-Smale complexes. In this talk we focus on the reconstruction of the dataset coherent with the simplified Morse-Smale complexes. This leads to simplified datasets in which the main topological features of the data are preserved. In comparison with previous related works, the novelty of our approach is to use techniques and results known from Computer Aided Geometric Design in order to avoid costly iterative optimization techniques. Following an introductory part on Morse-Smale complexes definition and simplification, we will explain the different steps involved in our method and illustrate them on several examples.