Download presentation
1
Shape Analysis and Retrieval
Extended Gaussian Images Notes courtesy of Funk et al., SIGGRAPH 2004
2
Extended Gaussian Image
Represent a model by a spherical function by binning surface normals Model Angular Bins EGI
3
Subdividing the Sphere
20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces All bins have the same area
4
Hierarchical Search 20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces
5
Hierarchical Search 20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces
6
Hierarchical Search 20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces
7
Hierarchical Search 20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces
8
Hierarchical Search 20 Faces Icosahedron 80 Faces 320 Faces 1280 Faces
9
Angular Parameterization
4x4 Bins 8x8 Bins 16x16 Bins 32x32 Bins Different bins have different areas
10
Angular Parameterization
Missing node where bins are small Different bins have different areas Smaller bins lose their votes
11
Angular Parameterization
Normalize bins by bin area
12
Extended Gaussian Image
Properties: Invertible for convex shapes 2D array of information Can be defined for most models Point Clouds Polygon Soups Closed Meshes Genus-0 Meshes Shape Spectrum
13
Extended Gaussian Image
Properties: Invertible for convex shapes 2D array of information Can be defined for most models Limitations: Too much information is lost Normals are sensitive to noise 3D Model EGI
14
Extended Gaussian Image
Properties: Invertible for convex shapes 2D array of information Can be defined for most models Limitations: Too much information is lost Normals are sensitive to noise Initial Model Noisy Model
15
Retrieval Results Princeton Shape Benchmark: (~900 models, ~90 classes) 14 biplanes 50 human bipeds 7 dogs 17 fish 16 swords 6 skulls 15 desk chairs 13 electric guitars
16
Retrieval Results Precision Recall
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.