Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001
DOI: 10.1109/cvpr.2001.990988
|View full text |Cite
|
Sign up to set email alerts
|

Abstract: This paper explores a view-based approach to recognize free-form objects in range images. We are using a set of local features that are easy to calculate and robust to partial occlusions. By combining those features in a multidimensional histogram, we can obtain highly discriminant classifiers without the need for segmentation. Recognition is pegomzed using either histogram matching or a probabilistic recognition algorithm. We compare the pelfonnance of both methods in the presence of occlusions and test the s… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
93
0

Publication Types

Select...
5
3
1

Relationship

0
9

Authors

Journals

citations
Cited by 151 publications
(94 citation statements)
references
References 18 publications
1
93
0
Order By: Relevance
“…3 and 4. The use of contextual or local descriptors is a well-established approach, which has been investigated thoroughly both for appearance-based keypoints in image data (see, e.g., [2,3,52]) and for 3D point clouds (see, e.g., [25,34,74]). …”
Section: Local Contextual Representationmentioning
confidence: 99%
“…3 and 4. The use of contextual or local descriptors is a well-established approach, which has been investigated thoroughly both for appearance-based keypoints in image data (see, e.g., [2,3,52]) and for 3D point clouds (see, e.g., [25,34,74]). …”
Section: Local Contextual Representationmentioning
confidence: 99%
“…The formal statistical method for assessing the dissimilarity between two probability functions is the χ 2 -test [8]. Another equivalent comparison measure to histograms is the intersection measurement, a similarity function which quantifies the common parts of two histograms.…”
Section: B Pairwise Region Correspondencementioning
confidence: 99%
“…There is a large number of dissimilarity functions to compare histograms, such as euclidean distance, quadratic form distance [3], statistical and probabilistic approaches [13], [8], among others. In this paper we restrict ourselves to statistical similarity functions.…”
Section: B Pairwise Region Correspondencementioning
confidence: 99%
“…One of the earliest related approaches, also evaluated on synthetic range images, was proposed in [17], where a histogram representation of a complete point cloud was built merging information of the pixel depths, normal orientations and surface curvatures. More recently, Flint et al [18] defined a descriptor for Hessian-based interest points by accumulating in a histogram the elevation difference of the normals estimated with different sized planes for all points in a support region.…”
Section: D Descriptorsmentioning
confidence: 99%