Home > Program > Tutorials > AM-04

ICPR2012 Tutorials AM-04
3D Shape Analysis and Retrieval – Recent Advances and Trends

Hamid Laga* and RyutarouOhbuchi**
*School of Mathematics and Statistics, University of South Australia, Australia
**Computer Science and Engineering Department, University of Yamanashi, Japan



In recent years, acquisition and modelling of 3D data has gained a significant boost due to the availability of commodity devices. Digital 3D shape models are becoming a key component in many industrial, entertainment and scientific sectors. Consequently, large collections of 3D data are nowadays available both in the public (e.g., on the Internet) as well as in private domains. Analysing, classifying, and querying such 3D data collections are becoming topics of increasing interest in the computer vision, pattern recognition, computer graphics and digital geometry processing communities. 3D shape analysis poses new challenges that are not existent in image and video analysis. The purpose of this tutorial is to introduce the foundation of this topic to the pattern recognition community, and overview the state-of-the-art techniques. The tutorial will start by introducing basic concepts such as 3D shape representations and shape descriptors, while outlining the major requirements and challenges. Then the tutorial looks at the fundamental problem of comparing shapes, where one seeks to design similarity measures that capture shape properties (ranging from geometry to semantics), and which are robust to different variabilities (such as non-rigid deformations). We will also discuss roles that machine learning plays in 3D shape analysis and retrieval. Then, we will review recent works on query specification for 3D retrieval. We conclude the tutorial with an overview of some (classical and non-classical) applications where 3D shape analysis plays a central role.


Course description

I. Introduction

  1. Motivation
     (why do we need to compare shapes)
  2. Taxonomy and problem matrix
    (Shape v.s. other properties, Global vs. partial similarity, intrinsic vs. extrinsic, geometry vs. semantics).

II. Rigid 3D shape analysis

  1. Introduction
  2. 3D shape representations
    (Depth maps, polygon soup, point cloud, watertight surfaces)
  3. Invariance properties
    (Global geometrical transformations, deformations, representations)
  4. Comparing rigid 3D shapes by global similarity
    (alignment and normalization, global shape descriptors).
  5. Similarity measures.
  6. Summary Introduction

III. Non-rigid 3D shape analysis

  1. Introduction
  2. Local geometrical features
    i. Basic local geometrical features.
    ii. Bag of Words (BoW) and Bag of Local descriptors.
  3. Manifold-based analysis of deformable shapes
    i. Diffusion geometry for isometry-invariant shape analysis.
    ii. A Riemannian framework for elastic shape analysis and shape statistics.
    iii. Comparison.
  4. Summary

IV. Learning

  1. Introduction
  2. Learning for better features
  3. Learning to capture semantics and intention
  4. Summary


V. Querying 3D shape databases

  1. Query by 3D example.
  2. Query by images, by depth images, and by sketch.
  3. Text-based and Context-based Search of 3D ModelsQuery by text and by example.


VI. Applications (15 min)

  1. 3D shape retrieval and classification.
  2. Archaeology and cultural heritage
  3. Biomedical application.
  4. 3D modeling.


VII. Conclusions and Wrap-up


VIII. Open discussions


Relevant References:

I. Descriptor-based 3D shape analysis
# Global descriptors

  • Shape Distributions
    R. Osada, T. Funkhouser, Bernard Chazelle, and David Dobkin, ACM TOG, 21(4): 807-832.

  • On Visual Similarity Based 3D Model Retrieval

  • Ding-Yun Chen, Xiao-Pei Tian, Yu-Te Shen and Ming Ouhyoung
    Computer Graphics Forum (EUROGRAPHICS'03), 22(3):223-232, Sept. 2003

  • Discriminative Spherical Wavelet Features for Content-based 3D Model Retrieval

  • Hamid Laga, Masayuki Nakajima, Kunihiro Chihara.
    International Journal on Shape Modeling, 13(1):51-72, 2007.

  • Rotation Invariant Spherical Harmonic Representation of 3D. Shape Descriptors

  • Michael Kazhdan, Thomas Funkhouser, and Szymon Rusinkiewicz
    Proc. 2003 Symposium on Geometry Processing, pp.156-164 (2003).

# Local descriptors

  • Recognizing Objects in Range Data Using Regional Point Descriptors.

  • Andrea Frome, Daniel Huber, Ravi Kolluri, Thomas Bulow, and Jitendra Malik.
    Proc. European Conference on Computer Vision (ECCV), May, 2004.

  • Robust Global Registration.

  • N. Gelfand, N. Mitra, L. Guibas and H. Pottmann. 
    Proc. 2005 Symposium on Geometry Processing, pp. 197-206, 2005.

  • Using spin images for efficient object recognition in cluttered 3d scenes.

  • Johnson, A. E., Hebert, M., IEEE Trans. Pattern Anal. Mach. Intell. 21, 433–449, 1999.

  • Salient local visual features for shape-based 3D model retrieval.

  • Ryutarou Ohbuchi, Kunio Osada, Takahiko Furuya, Tomohisa Banno,
    Proc. IEEE Shape Modeling International 2008: 93-10.

  • Hough Transform and 3D SURF for robust three dimensional classification

  • Jan Knopp, Mukta Prasad, Geert Willems, Radu Timofte, and Luc Van Gool,
    Proc. ECCV 2010, pp.589-602.

  • Scale Invariant Features for 3D Mesh Models

  • Tal Darom and Yosi Keller, to appear, IEEE Trans. Image Processing, 21(5): 2758-2769 (2012).

II. Shape similarity
# Bag of words

  • Shape Google: geometric words and expressions for invariant shape retrieval

  • M. Bronstein, M. M. Bronstein, M. Ovsjanikov, L. J. Guibas,
    ACM Trans. Graphics (TOG), Vol. 30/1, pp. 1-20, January 2011

  • # Isometry invariant shape analysis

    • Pose oblivious Shape Signature

    • Ran Gal, Ariel Shamir, Daniel Cohen-Or.
      IEEE Transactions on Visualization and Computer Graphics 2007.

    • Non-Rigid Spectral Correspondence of Triangle Meshes

    • Varun Jain, Hao Zhang, and Oliver van Kaick. Int. Journal on Shape Modeling (2007).

    • A Spectral Approach to Shape-Based Retrieval of Articulated 3D Models

    • Varun Jain and Hao Zhang, Computer-Aided Design, Vol. 39, Issue 5, pp. 398-407, 2007.

    # Diffusion Geometry

    • A Concise and Provably Informative Multi-scale Signature Based on Heat Diffusion,

    • Jian Sun, Maks Ovsjanikov, and Leonidas Guibas,

    • Proc. Eurographics Symposium on Geometry Processing (SGP) 2009.

    • Laplace-Beltrami spectra as "Shape-DNA" of surfaces and solids.

    • M. Reuter, F.-E. Wolter and N. Peinecke. Computer-Aided Design 38 (4), pp.342-366, 2006.

    • Intrinsic Shape Context Descriptors for Deformable Shapes

    • Iasonas Kokkinos, Michael Bronstein, Roee Litman, Alexander Bronstein. CVPR 2012.

    # Elastic shape analysis

    • Elastic Geodesic Paths in Shape Spaces of Parameterized Surfaces

    • S. Kurtek, E. Klassen, J. Gore, Z. Ding, and A. Srivastava
      IEEE Transactions on Pattern Analysis and Machine Intelligence (2011).

    III. Some semantics

    • Contextual Part Analogies in 3D Objects

    • Shapira, S. Shalom, A. Shamir, D. Cohen-Or, and H. Zhang.
      Int. J. Comput. Vision 89, pp. 309-326 (2009)

    IV. Supervised / unsupervised Learning for 3D shape analysis and retrieval

    • Semantic-oriented 3D shape retrieval using relevance feedback
      George Leifman, Ron Meir and Ayellet Tal.

    • The Visual Computer, 21 (8-10) (2005), pp. 865-875

    • Unsupervised learning from a corpus for shape-based 3D model retrieval.

    • Ryutarou Ohbuchi, Jun Kobayashi: Multimedia Information Retrieval 2006: 163-172

    • Distinctive Regions of 3D Surfaces

    • Philip Shilane and Thomas Funkhouser. ACM Transactions on Graphics, 26(2), June 2007

    • Automatic Selection of Best Views of 3D Shapes

    • Hamid Laga,The Visual Computer 2011.

    V. Querying 3D shape databases

    • How Do Humans Sketch Objects?

    • Mathias Eitz, James Hays, Marc Alexa. Transactions on Graphics, Proc. SIGGRAPH 2012

    • Context-based Search of 3D Models

    • M. Fisher, P. Hanrahan. ACM Transactions on Graphics (Proc. SIGGRAPH Asia 2010).

    • Autotagging to Improve Text Search for 3D Models

    • Corey Goldfeder, Peter K. Allen, Joint Conference on Digital Libraries 2008.

    • Characterizing Structural Relationships in Scenes Using Graph Kernels
      M. Fisher, M. Savva, P. Hanrahan. SIGGRAPH 2011.

    • 3D-Model Search Engine from Photos

    • T.F. Ansary, J-P. Vandeborre, Mohamed Daoudi. Proc. ACM CIVR 2007, pp. 89-92

    VII. Applications:

    • Modeling by example

    • T. Funkhouser, M. Kazhdan, P. Shilane, P. Min, A. Tak. S. Rusinkiewicz, and D. Dobkin.
      ACM TOG (23)3:652-663. (Proc. SIGGRAPH 2004), 2004.

    • Reassembling Fractured Objects by Geometric Matching

    • Qi-Xing Huang. Simon Flöry, Nathsha Gelfand, Michael Hofer and Helmut Pottmann.
      ACM TOG 25(3):569-578,(Proc. SIGGRAPH 2006), 2006.

    • Exploring Collections of 3D Models using Fuzzy Correspondences

    • Vladimir G. Kim, Wilmot Li, Niloy Mitra, Stephen DiVerdi, Thomas Funkhouser
      ACM TOG 31(4), (Proc. SIGGRAPH 2012), 2012.

    • Artwork 3D model database indexing and classification
      Slvie Philipp-Foliquet, Michael Jordan, Laurent Najman, Jean Cousty, Pattern Recognition, 44(3), March 2011, pp.588-597.

    • Similarity Based Object Retrieval of Composite Neuronal Structures
      F. Shulze, M. Trapp, K. Buhler, T. Liu, B. Dickson, Proc. Eurographics 3DOR 2012.


    VIII. Datasets and benchmarks

    • The Princeton Shape Benchmark:

    • Shape Retrieval Evaluation Context (SHREC):

    • The TOSCA datasets:

    • Shape Retrieval Contest:


    IX. Other (related / similar) tutorials

    • Shape-based retrieval and analysis of 3D models

    • Siggraph 2004 course,

    • Diffusion Geometry in 3D Shape Analysis

    • M. Bronstein, U. Castellani, and A. Bronstein
      Eurographics 2012 Tutorial:

    • Spectral Geometry Processing

    • Bruno Levy and Richard Hao Zhang. ACM SIGGRAPH Course Notes (2010).



    About Lecturer:

    Hamid Laga is currently a senior researcher at the School of Mathematics and Statistics of the University of South Australia. Previously he has been an Associate Professor at the Institut Telecom (France), Assistant Professor at Tokyo Institute of Technology (Japan), and a JSPS post-doctoral fellow at Nara Institute of Science and Technology (Japan). Hamid received his PhD degree in 2006 from Tokyo Institute of Technology (Japan) on 3D shape analysis and retrieval. He published several papers in this field. His work on shape descriptors for 3D retrieval received the best paper award at IEEE Conference on Shape Modeling and Applications (SMI) 2006. He co-chaired the Eurographics Workshop on 3D Object Retrieval (3DOR) 2011. His recent research interests include semantic analysis of shapes and shape statistics. Further details are available at:


    Ryutarou Ohbuchi is currently a Professor at Department of Computer Science and Engineering, School of Engineering, University of Yamanashi (Japan). Previously, he was a senior research at IBM Research Tokyo. Ohbuchi received his Ph.D degree in 1994 from University of North Carolina at Chapel Hill (U.S.A) on augmented reality display for medical applications. While at IBM Research, in 1997, he and his colleagues pioneered digital watermarking of 3D mesh models. His current research interests include shape-based 3D model retrieval. Further details are available at:


    > Back to List of Tutorials


    Archive: Call for Tutorial Proposals



    (C) 2010-2012 ICPR2012 Organizing Committee. All rights reserved.