Unsupervised learning of object features from video sequences

Marius Leordeanu; Robert Collins

doi:10.1109/CVPR.2005.359

Unsupervised learning of object features from video sequences

Marius Leordeanu, Robert Collins

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

36 Scopus citations

Abstract

We develop an efficient algorithm for unsupervised learning of object models as constellations of features, from low resolution video sequences. The input images typically contain single or multiple objects that change in pose, scale and degree of occlusion. Also, the objects can move significantly between consecutive frames. The content of an input sequence is unlabeled so the learner has to cluster the data based on the data's implicit coherence over time and space. Our approach takes advantage of the dependent pairwise co-occurrences of objects' features within local neighborhoods vs. the independent behavior of unrelated features. We couple or decouple pairs of features based on a probabilistic interpretation of their pairwise statistics and then extract objects as connected components of features.

Original language	English (US)
Title of host publication	Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005
Publisher	IEEE Computer Society
Pages	1142-1149
Number of pages	8
ISBN (Print)	0769523722, 9780769523729
DOIs	https://doi.org/10.1109/CVPR.2005.359
State	Published - 2005
Event	2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005 - San Diego, CA, United States Duration: Jun 20 2005 → Jun 25 2005

Publication series

Name	Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005
Volume	I

Other

Other	2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005
Country/Territory	United States
City	San Diego, CA
Period	6/20/05 → 6/25/05

All Science Journal Classification (ASJC) codes

General Engineering

Access to Document

10.1109/CVPR.2005.359

Cite this

Leordeanu, M., & Collins, R. (2005). Unsupervised learning of object features from video sequences. In Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005 (pp. 1142-1149). Article 1467395 (Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005; Vol. I). IEEE Computer Society. https://doi.org/10.1109/CVPR.2005.359

@inproceedings{ec652667b3e04c99b738edb1dcb5f138,

title = "Unsupervised learning of object features from video sequences",

abstract = "We develop an efficient algorithm for unsupervised learning of object models as constellations of features, from low resolution video sequences. The input images typically contain single or multiple objects that change in pose, scale and degree of occlusion. Also, the objects can move significantly between consecutive frames. The content of an input sequence is unlabeled so the learner has to cluster the data based on the data's implicit coherence over time and space. Our approach takes advantage of the dependent pairwise co-occurrences of objects' features within local neighborhoods vs. the independent behavior of unrelated features. We couple or decouple pairs of features based on a probabilistic interpretation of their pairwise statistics and then extract objects as connected components of features.",

author = "Marius Leordeanu and Robert Collins",

year = "2005",

doi = "10.1109/CVPR.2005.359",

language = "English (US)",

isbn = "0769523722",

series = "Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005",

publisher = "IEEE Computer Society",

pages = "1142--1149",

booktitle = "Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005",

address = "United States",

note = "2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005 ; Conference date: 20-06-2005 Through 25-06-2005",

}

Leordeanu, M & Collins, R 2005, Unsupervised learning of object features from video sequences. in Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005., 1467395, Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005, vol. I, IEEE Computer Society, pp. 1142-1149, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005, San Diego, CA, United States, 6/20/05. https://doi.org/10.1109/CVPR.2005.359

Unsupervised learning of object features from video sequences. / Leordeanu, Marius; Collins, Robert.
Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005. IEEE Computer Society, 2005. p. 1142-1149 1467395 (Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005; Vol. I).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Unsupervised learning of object features from video sequences

AU - Leordeanu, Marius

AU - Collins, Robert

PY - 2005

Y1 - 2005

N2 - We develop an efficient algorithm for unsupervised learning of object models as constellations of features, from low resolution video sequences. The input images typically contain single or multiple objects that change in pose, scale and degree of occlusion. Also, the objects can move significantly between consecutive frames. The content of an input sequence is unlabeled so the learner has to cluster the data based on the data's implicit coherence over time and space. Our approach takes advantage of the dependent pairwise co-occurrences of objects' features within local neighborhoods vs. the independent behavior of unrelated features. We couple or decouple pairs of features based on a probabilistic interpretation of their pairwise statistics and then extract objects as connected components of features.

AB - We develop an efficient algorithm for unsupervised learning of object models as constellations of features, from low resolution video sequences. The input images typically contain single or multiple objects that change in pose, scale and degree of occlusion. Also, the objects can move significantly between consecutive frames. The content of an input sequence is unlabeled so the learner has to cluster the data based on the data's implicit coherence over time and space. Our approach takes advantage of the dependent pairwise co-occurrences of objects' features within local neighborhoods vs. the independent behavior of unrelated features. We couple or decouple pairs of features based on a probabilistic interpretation of their pairwise statistics and then extract objects as connected components of features.

UR - http://www.scopus.com/inward/record.url?scp=33745146196&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745146196&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2005.359

DO - 10.1109/CVPR.2005.359

M3 - Conference contribution

AN - SCOPUS:33745146196

SN - 0769523722

SN - 9780769523729

T3 - Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005

SP - 1142

EP - 1149

BT - Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005

PB - IEEE Computer Society

T2 - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005

Y2 - 20 June 2005 through 25 June 2005

ER -

Leordeanu M, Collins R. Unsupervised learning of object features from video sequences. In Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005. IEEE Computer Society. 2005. p. 1142-1149. 1467395. (Proceedings - 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005). doi: 10.1109/CVPR.2005.359

Unsupervised learning of object features from video sequences

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this