1990
DOI: 10.1007/bf00127812
|View full text |Cite
|
Sign up to set email alerts
|

Automatic extraction of deformable part models

Abstract: I describe a system that first segments an image into parts, and then fits deformable models to range data associated with the image. Models are represented by using modal dynamics applied to volumetric primitives, which significantly improves the computational complexity of both model recovery and subsequent processing. The segmentation procedure uses two simple mechanisms: a filtering operation to produce a large set of potential object parts, followed by a quadratic integer optimization that searches among … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
88
0
1

Year Published

1994
1994
2007
2007

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 201 publications
(89 citation statements)
references
References 29 publications
0
88
0
1
Order By: Relevance
“…They have been used for 2-D surface registration [1] and were rapidly extended to various 3-D formulations [3,13,12]. To reduce the dimensionality of the models, linearity assumptions have also been made through modal analysis [14,3,5]. Even though the physics-based approach has been extremely successful, it implies some knowledge of the pseudo-physical properties of the surface, which may not be available.…”
Section: Deformable Shape Recoverymentioning
confidence: 99%
See 1 more Smart Citation
“…They have been used for 2-D surface registration [1] and were rapidly extended to various 3-D formulations [3,13,12]. To reduce the dimensionality of the models, linearity assumptions have also been made through modal analysis [14,3,5]. Even though the physics-based approach has been extremely successful, it implies some knowledge of the pseudo-physical properties of the surface, which may not be available.…”
Section: Deformable Shape Recoverymentioning
confidence: 99%
“…Structure from motion techniques [18,10,22] and physics-based models [13,14,3,5,12] have been proposed to retrieve the deformations of non-rigid surfaces. However, these approaches typically incorporate regularization terms or make linearity assumptions that constrain the surface to deform smoothly.…”
Section: Introductionmentioning
confidence: 99%
“…Recognition is based on an indexing process that relies on the structural composition of the object. Examples for such an approach can be found in [5,[12][13][14]32,38,39]. Although these approaches are more generic and robust than the geometric model based schemes, they are still confined to recognition problems in which the shape of the objects to be recognized is known.…”
Section: Related Workmentioning
confidence: 99%
“…Most of those approaches are applied to range data only (e.g., Solina and Bajcsy 1990;Gupta 1991), while Pentland (1990) describes a two-stage algorithm to fit superquadrics to image data. In the first stage, he segments the image using a filtering operation to produce a large set of potential object "parts", followed by a quadratic optimization procedure that searches among these part hypotheses to produce a maximum likelihood estimate of the image's part structure.…”
Section: Related Workmentioning
confidence: 99%