View-based object recognition guided by intelligent visual exploration

Qiang Lu; Peter D. Scott

View-based object recognition guided by intelligent visual exploration

Qiang Lu
, Peter D. Scott

Department of Computer Science and Engineering

SUNY Buffalo

Research output: Contribution to conference › Paper › peer-review

Abstract

We present an active 3-D object recognition scheme in which, when an ambiguous object is viewed, a new imaging viewpoint is computed to efficiently reduce the predicted object class ambiguity. The camera moves toward that viewpoint and a new image of the object is acquired. The data from the set of images of the object acquired to-date is combined, and the consistent hypothesis set reduced as the number of these views increases, until recognition is achieved. Detailed approaches are presented and discussed along with preliminary experimental results.

Original language	English
Pages	957-962
Number of pages	6
State	Published - 2003
Event	Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference - St. Louis, MO., United States Duration: Nov 2 2003 → Nov 5 2003

Conference

Conference	Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference
Country/Territory	United States
City	St. Louis, MO.
Period	11/2/03 → 11/5/03

Cite this

@conference{5ba1097505e54490a4f643f30db61770,

title = "View-based object recognition guided by intelligent visual exploration",

abstract = "We present an active 3-D object recognition scheme in which, when an ambiguous object is viewed, a new imaging viewpoint is computed to efficiently reduce the predicted object class ambiguity. The camera moves toward that viewpoint and a new image of the object is acquired. The data from the set of images of the object acquired to-date is combined, and the consistent hypothesis set reduced as the number of these views increases, until recognition is achieved. Detailed approaches are presented and discussed along with preliminary experimental results.",

author = "Qiang Lu and Scott, \{Peter D.\}",

year = "2003",

language = "English",

pages = "957--962",

note = "Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference ; Conference date: 02-11-2003 Through 05-11-2003",

}

Lu, Q & Scott, PD 2003, 'View-based object recognition guided by intelligent visual exploration', Paper presented at Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference, St. Louis, MO., United States, 11/2/03 - 11/5/03 pp. 957-962.

View-based object recognition guided by intelligent visual exploration. / Lu, Qiang; Scott, Peter D.
2003. 957-962 Paper presented at Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference, St. Louis, MO., United States.

Research output: Contribution to conference › Paper › peer-review

TY - CONF

T1 - View-based object recognition guided by intelligent visual exploration

AU - Lu, Qiang

AU - Scott, Peter D.

PY - 2003

Y1 - 2003

N2 - We present an active 3-D object recognition scheme in which, when an ambiguous object is viewed, a new imaging viewpoint is computed to efficiently reduce the predicted object class ambiguity. The camera moves toward that viewpoint and a new image of the object is acquired. The data from the set of images of the object acquired to-date is combined, and the consistent hypothesis set reduced as the number of these views increases, until recognition is achieved. Detailed approaches are presented and discussed along with preliminary experimental results.

AB - We present an active 3-D object recognition scheme in which, when an ambiguous object is viewed, a new imaging viewpoint is computed to efficiently reduce the predicted object class ambiguity. The camera moves toward that viewpoint and a new image of the object is acquired. The data from the set of images of the object acquired to-date is combined, and the consistent hypothesis set reduced as the number of these views increases, until recognition is achieved. Detailed approaches are presented and discussed along with preliminary experimental results.

UR - https://www.scopus.com/pages/publications/2442519509

M3 - Paper

AN - SCOPUS:2442519509

SP - 957

EP - 962

T2 - Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life - Proceedings of the Artificial Neural Networks in Engineering Conference

Y2 - 2 November 2003 through 5 November 2003

ER -

View-based object recognition guided by intelligent visual exploration

Abstract

Conference

Other files and links

Fingerprint

Cite this