|
|
3D Shape Comparison
This is a journal while I am doing research on 3D Shape Comparison. I will periodically put something on. Firstly, I put papers with annotations by myself for futher reference.
Terminology
Thought
Reference| 2003/10/28 | | Topology Matching for Fully Automatic Similarity Estimation of 3D Shapes | | Masaki Hilaga, Yoshihisa Shinagawa, Taku Kohmura, Tosiyasu L. Kunii | | SIGGRAPH 2001 |
| | The constructed Reeb graph is invariant to translation and rotation, robust against connectivity changes caused by simplification, subdivision, and remesh, and resistant against noise and certain changes due to deformation. | | Reeb graph is defined by a function. It is used to do decomposition and retrieve topology information from shapes. | | matching algorithm := similarity + topology. (for each node in the Reeb graph) | | similarity := area ratio + length. |
| 2003/10/28 | | A Novel Method for 3D Surface Mesh Segmentation | | Thitiwan Srinark, Chandra Kambhamettu | | 6th IASTED International Conference on Computers, Graphics, and Imaging, Honolulu, Hawaii, USA, August 13-15, 2003 |
| | The surface type of a point can be classified using Gaussian curvature and mean curvature using Besl and Jain[Segmentation through variable-order surface fitting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988] | | It can determine several characters of surface mesh: peak, ridge, saddle ridge, flat, minimal surface, pit, valley, saddle valley. |
| 2003/11/04 | Matching 3D Models with Shape Distributions
Robert Osada, Thomas Funkhouser, Bernard Chazelle, David Dobkin
Shape Modeling International, 2001 |
| | Reduce the shape matching problem to the comparison of probability distributions. | | Be applied as a pre-classifier in an object recognition/retrieval system. | | applicable for missing, wrongly-oriented, intersecting, disjoint and overlapping polygons. | | Shape functions: D1, D2, D3, D4, A3... | | The shape distribution is too large such that stochastic methods are employed. | | Must concerns: sampling density. | | This paper focuses on analyses of shape functions. |
| 2003/11/04 | Efficient Multiple Model Recognition in Cluttered 3-D Scenes
''Andrew Edie Johnson, Martial Hebert
Proc. Computer Vision and Pattern Recognition (CVPR `98) |
| | Local descriptor, Spin-images | | Recognition of multiple objects in scenes containing clutter and occlusion. | | Spin-images of small support are robust to clutter and occlusion while spin-images of large support are highly discriminating. | | Occlusion := the ratio of remaining area to the whole area of a model in the scene. | | Clutter := the ration of the additional volume to the original volume of a model in the scene. | Alignment by spin-images.
Matching by Iterative Closest Point(ICP).
Compression by Principal Component Analysis(PCA). |
|