Reconciling landmarks and level sets: geometric shape warping and matching using generalized gradients and correspondence-augmented implicit representations

نویسندگان

  • Pierre Maurel
  • Renaud Keriven
  • Olivier Faugeras
چکیده

Shape warping is a key problem in statistical shape analysis. This paper proposes a framework for geometric shape warping based on both shape distances and landmarks. Taking advantage of the recently proposed spatially coherent flows, our method is mathematically well-posed and uses only intrinsic shape information, namely some similarity measure between shapes and the correspondence of landmarks provided on the shape surface. No extrinsic quantity is considered, neither a diffeomorphism of the embedding space nor point correspondences in this space. Thanks to a recent extension of the level set method allowing point tracking and tangential velocities, our method is compatible with implicit representations. Moreover, a matching between shape surfaces is provided at no additional cost. Although some recent work deals with implicit representations and landmarks, it is, to our knowledge, the first time that landmarks and shape distances are reconciled in a pure geometric level set framework. The feasibility of the method is demonstrated with twoand three-dimensional examples. Combining shape distance and landmarks, our approach reveals to need only a small number of landmarks to obtain improvements on both warping and matching. Key-words: Shape metrics, distance fonctions, level set, shape warping, shape matching, landmarks Points d’amer et méthode des ensembles de niveaux: Déformation de formes et mise en correspondance Résumé : La déformation de formes est un problème important de l’analyse statistique des formes. Nous proposons dans cet article un cadre permettant de déformer une forme en une autre. Notre méthode utilise à la fois des distances entre formes et des points d’amer (ou landmarks). En utilisant la technique récemment proposée des gradients avec cohérence spatiale, notre travail est mathématiquement bien posé et utilise seulement l’information intrinsèque des formes: une mesure de la similarité entre formes et la correspondance entre points d’amer fournie sur la surface de la forme. Aucune quantité extrinsèque n’est considérée: nul besoin d’un difféomorphisme de l’espace dans lequel la forme est plongée, ni de correspondances entre les points de cet espace. Grâce à la récente extension de la méthode des ensembles de niveaux permettant le suivi de point ainsi que l’introduction de vitesses tangentielles, notre méthode est compatible avec une représentation implicite des formes. De plus un appariement entre les formes est fourni sans coût supplémentaire. Bien que certains récents travaux traitent du sujet de représentation implicite et de points d’amer, il s’agit, à notre connaissance, de la première fois que les points d’amer et les distances entre formes sont réunis dans un cadre purement géométrique et implicite. La faisabilité de cette méthode est démontrée par des exemples en deux et trois dimensions. Associant distance entre forme et points d’amer, notre approche ne nécessite qu’un petit nombre de points d’amer pour obtenir une amélioration à la fois de la déformation de la forme mais aussi de la mise en correspondance. Mots-clés : Métriques de formes, fonctions distances, ensembles de niveaux, déformation de formes, appariement de formes, points d’amer Landmarks and level sets: shape warping and matching 3

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Shape Modeling Using Automatic Landmarking

This paper describes a novel approach to automatically recover accurate correspondence over various shapes. In order to detect the features points with the capability in capturing the characteristics of an individual shape, we propose to calculate the skeletal representation for the shape curve through the medial axis transform. Employing this shape descriptor, mathematical landmarks are automa...

متن کامل

Body shape comparison of Big-head carp with two variants of silver carp using geometric morphometric techniques

In order to investigate the differences in the shape of the big-head and two morphotypes of Phytophagous fish by geometric morphometric technique, 30 samples of each group, with a mean standard length of 25 ± 3 cm, were prepared. After fish were anesthetized by 1% Clove powder extract and stabilized in formalin buffer 4%, 6-megapixel digital images of the left profile of the specimens were prep...

متن کامل

Phenotypic plasticity of the body shape in wild and farmed common carp (Cyprinus carpio) using geometric morphometric method

The aim of this study was to investigate the body shape and morphological comparison between the populations of wild (marine) and farmed common carp using geometric morphology. For this purpose, 65 pieces of fish including 23 wild common carp and 42 farmed common carp were sampled. The samples were then photographed and 16 landmarks were identified on the images. Procrast analysis was used in o...

متن کامل

Structure-oriented contour representation and matching for engineering shapes

Conventional shape matching for engineering models primarily considers rigid shape similarity. They do not seek global shape similarity while considering large local deformations. However, engineering models created by some parametric-based design can involve large parametric changes. As a result, they do not share similarity in their global shape. Hence our goal is to develop shape representat...

متن کامل

Recursive scan-matching SLAM

This paper presents Scan-SLAM, a new generalization of simultaneous localization and mapping (SLAM). SLAM implementations based on extended Kalman filter (EKF) data fusion have traditionally relied on simple geometric models for defining landmarks. This limits EKF-SLAM to environments suited to such models and tends to discard much potentially useful data. The approach presented in this paper i...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2005