Results 11  20
of
79
A Bag of Words Approach for 3D Object Categorization
"... Abstract. In this paper we propose a novel framework for 3D object categorization. The object is modeled it in terms of its subparts as an histogram of 3D visual word occurrences. We introduce an effective method for hierarchical 3D object segmentation driven by the minima rule that combines spectr ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
(Show Context)
Abstract. In this paper we propose a novel framework for 3D object categorization. The object is modeled it in terms of its subparts as an histogram of 3D visual word occurrences. We introduce an effective method for hierarchical 3D object segmentation driven by the minima rule that combines spectral clustering – for the selection of seedregions – with region growing based on fast marching. The front propagation is driven by local geometry features, namely the Shape Index. Finally, after the coding of each object according to the BagofWords paradigm, a Support Vector Machine is learnt to classify different objects categories. Several examples on two different datasets are shown which evidence the effectiveness of the proposed framework. 1
Computing Teichmüller Shape Space
 SUBMITTED TO IEEE TVCG
"... Shape indexing, classification, and retrieval are fundamental problems in computer graphics. This work introduces a novel method for surface indexing and classification based on Teichmüller theory. Two surfaces are conformal equivalent, if there exists a bijective anglepreserving map between them. ..."
Abstract

Cited by 6 (3 self)
 Add to MetaCart
Shape indexing, classification, and retrieval are fundamental problems in computer graphics. This work introduces a novel method for surface indexing and classification based on Teichmüller theory. Two surfaces are conformal equivalent, if there exists a bijective anglepreserving map between them. The Teichmüller space for surfaces with the same topology is a finite dimensional manifold, where each point represents a conformal equivalence class, and the conformal map is homotopic to Identity. A curve in the Teichmüller space represents a deformation process from one class to the other. In this work, we apply Teichmüller space coordinates as shape descriptors, which are succinct, discriminating and intrinsic, invariant under the rigid motions and scalings, insensitive to resolutions. Furthermore, the method has solid theoretic foundation, and the computation of Teichmüller coordinates is practical, stable and efficient. The algorithms for the Teichmüller coordinates of surfaces with positive or zero Euler numbers have been studied before. This work focuses on the surfaces with negative Euler numbers, which have a unique conformal Riemannian metric with −1 Gaussian curvature. The coordinates which we will compute are the lengths of a special set of geodesics under this special metric. The metric can be obtained by the curvature flow algorithm, the geodesics can be calculated using algebraic topological method. We tested our method extensively for indexing and comparison of about one hundred of surfaces with various topologies, geometries and resolutions. The experimental results show the efficacy and efficiency of the length coordinate of the Teichmüller space.
A study of shape similarity for temporal surface sequences of people
 In appear in Proceedings of 3DIM 2007
, 2007
"... The problem of 3D shape matching is typically restricted to static objects to classify similarity for shape retrieval. In this paper we consider 3D shape matching in temporal sequences where the goal is instead to find similar shapes for a single timevarying object, here the human body. Localfeatur ..."
Abstract

Cited by 4 (3 self)
 Add to MetaCart
(Show Context)
The problem of 3D shape matching is typically restricted to static objects to classify similarity for shape retrieval. In this paper we consider 3D shape matching in temporal sequences where the goal is instead to find similar shapes for a single timevarying object, here the human body. Localfeature distribution descriptors are adopted to provide a rich object description that is invariant to changes in surface topology. Two contributions are made, (i) a comparison of descriptors for shape similarity in temporal sequences of a dynamic freeform object and (ii) a quantitative evaluation based on the ReceiverOperator Characteristic (ROC) curve for the descriptors using a groundtruth data set for synthetic motion sequences. Shape Distribution [25], Spin Image [15], Shape Histogram [1] and Spherical Harmonic [17] descriptors are compared. The highest performance is obtained by volumesampling shapehistogram descriptors. The descriptors also demonstrate relative insensitivity to parameter setting. The application is demonstrated in captured sequences of 3D human surface motion. 1.
Analogical recognition of shape and structure in design drawings. Artificial Intelligence for Engineering Design, Analysis and Manufacturing 22:117–128
"... We describe a method for constructing a structural model of an unlabeled target twodimensional line drawing by analogy to a known source model of a drawing with similar structure. The source case is represented as a schema that contains its line drawing and its structural model represented at multi ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
We describe a method for constructing a structural model of an unlabeled target twodimensional line drawing by analogy to a known source model of a drawing with similar structure. The source case is represented as a schema that contains its line drawing and its structural model represented at multiple levels of abstraction: the lines and intersections in the drawing, the shapes, the structural components, and connections of the device are depicted in the drawing. Given a target drawing and a relevant source case, our method of compositional analogy first constructs a representation of the lines and the intersections in the target drawing, then uses the mappings at the level of line intersections to transfer the shape representations from the source case to the target; next, it uses the mappings at the level of shapes to transfer the full structural model of the depicted system from the source to the target.
Content based 3d shape retrieval, a survey of state of the art
 Computer Science Ph.D. program 2nd Exam Part 1, http://web.cs.gc.cuny.edu/ ∼ iicke/academic/survey.pdf,2004
"... Large databases of 3dimensional (3D) data are becoming available on the Internet, and in various domains such as Computer Aided Design (CAD), Molecular Biology (3D Protein Models), Computer Graphics, Medicine and Archeology to name a few. As the number and variety of the 3D models(3D shapes) contin ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
(Show Context)
Large databases of 3dimensional (3D) data are becoming available on the Internet, and in various domains such as Computer Aided Design (CAD), Molecular Biology (3D Protein Models), Computer Graphics, Medicine and Archeology to name a few. As the number and variety of the 3D models(3D shapes) continue to grow, there has been an increasing interest in applications to help people search in these large databases. Even though each model might have a file name or other textual data associated with it, most of the time such information will not be enough to fully describe what the model actually is. The solution is to use the content, namely the shape. In this paper, we give an overview of the recent research on Content Based 3D Shape Retrieval. First, general shape similarity and matching concepts are given. A review of the literature on 3D shape matching research within the content based retrieval framework is the main part of this paper. Then the issues affecting the shape retrieval performance issues are discussed. 1
Temperature distribution descriptor for robust 3D shape retrieval
 IN COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW), 2011 IEEE COMPUTER SOCIETY CONFERENCE ON
, 2011
"... Recent developments in acquisition techniques are resulting in a very rapid growth of the number of available three dimensional (3D) models across areas as diverse as engineering, medicine and biology. It is therefore of great interest to develop the efficient shape retrieval engines that, given a q ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
Recent developments in acquisition techniques are resulting in a very rapid growth of the number of available three dimensional (3D) models across areas as diverse as engineering, medicine and biology. It is therefore of great interest to develop the efficient shape retrieval engines that, given a query object, return similar 3D objects. The performance of a shape retrieval engine is ultimately determined by the quality and characteristics of the shape descriptor used for shape representation. In this paper, we develop a novel shape descriptor, called temperature distribution (TD) descriptor, which is capable of exploring the intrinsic geometric features on the shape. It intuitively interprets the shape in an isometricallyinvariant, shapeaware, noise and small topological changes insensitive way. TD descriptor is driven by by heat kernel. The TD descriptor understands the shape by evaluating the surface temperature distribution evolution with time after applying unit heat at each vertex. The TD descriptor is represented in a concise form of a one dimensional (1D) histogram, and captures enough information to robustly handle the shape matching and retrieval process. Experimental results demonstrate the effectiveness of TD descriptor within applications of 3D shape matching and searching for the models at different poses and various noise levels.
Improving Generalization for 3D Object Categorization with Global Structure Histograms
"... Abstract — We propose a new object descriptor for three dimensional data named the Global Structure Histogram (GSH). The GSH encodes the structure of a local feature response on a coarse global scale, providing a beneficial tradeoff between generalization and discrimination. Encoding the structural ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
(Show Context)
Abstract — We propose a new object descriptor for three dimensional data named the Global Structure Histogram (GSH). The GSH encodes the structure of a local feature response on a coarse global scale, providing a beneficial tradeoff between generalization and discrimination. Encoding the structural characteristics of an object allows us to retain low local variations while keeping the benefit of global representativeness. In an extensive experimental evaluation, we applied the framework to categorybased object classification in realistic scenarios. We show results obtained by combining the GSH with several different local shape representations, and we demonstrate significant improvements to other stateoftheart global descriptors. I.
Computing FenchelNielsen Coordinates in Teichmüller Shape Space
 COMMUNICATIONS IN INFORMATION AND SYSTEM
, 2009
"... Teichmüller shape space is a finite dimensional Riemannian manifold, where each point represents a class of surfaces, which are conformally equivalent, and a path represents a deformation process from one shape to the other. Two surfaces in the real world correspond to the same point in the Teichmül ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
(Show Context)
Teichmüller shape space is a finite dimensional Riemannian manifold, where each point represents a class of surfaces, which are conformally equivalent, and a path represents a deformation process from one shape to the other. Two surfaces in the real world correspond to the same point in the Teichmüller space, only if they can be conformally mapped to each other. Teichmüller shape space can be used for surface classification purpose in shape modeling. This work focuses on the computation of the coordinates of high genus surfaces in the Teichmüller space. The coordinates are called as FenchelNielsen coordinates. The main idea is to deform the surface conformally using surface Ricci flow, such that the Gaussian curvature is −1 everywhere. The surface is decomposed to several pairs of hyperbolic pants. Each pair of pants is a genus zero surface with three boundaries, equipped with hyperbolic metric. Furthermore, all the boundaries are geodesics. Each pair of hyperbolic pants can be uniquely described by the lengths of its boundaries. The way of gluing different pairs of pants can be represented by the twisting angles between two adjacent pairs of pants which share a common boundary. The algorithms are based on Teichmüller space theory in conformal geometry, and they utilize the discrete surface Ricci flow. Most computations are carried out using hyperbolic geometry. The method is automatic, rigorous and efficient. The Teichmüller shape space coordinates can be used for surface classification and indexing. Experimental results on surfaces acquired from real world showed the practical value of the method for geometric database indexing, shape comparison and classification.
Comparison of dimension reduction methods for databaseadaptive 3D model retrieval
 Adaptive Multimedia Retrieval (AMR) 2007
, 2007
"... Abstract. Distance measures, along with shape features, are the most critical components in a shapebased 3D model retrieval system. Given a shape feature, an optimal distance measure will vary per query, per user, or per database. No single, fixed distance measure would be satisfactory all the time ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
Abstract. Distance measures, along with shape features, are the most critical components in a shapebased 3D model retrieval system. Given a shape feature, an optimal distance measure will vary per query, per user, or per database. No single, fixed distance measure would be satisfactory all the time. This paper focuses on a method to adapt distance measure to the database to be queried by using learningbased dimension reduction algorithms. We experimentally compare six such dimension reduction algorithms, both linear and nonlinear, for their efficacy in the context of shapebased 3D model retrieval. We tested the efficacy of these methods by applying them to five global shape features. Among the dimension reduction methods we tested, nonlinear manifold learning algorithms performed better than the other, e.g. linear algorithms such as principal component analysis. Performance of the best performing combination is roughly the same as the top finisher in the SHREC 2006 contest. 1.