Human face structure estimation from multiple images using the 2D affine space

Kuntal Sengupta, Jun Ohya

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In this paper we present an algorithm to estimate the human face structure. The input to the algorithm is not limited to an image sequence of a human head under rigid motion. It can be snapshots of the human face taken by the same or di erent cameras, over di erent periods of time. Since the depth variation of the human face is not very large, we use the a ne camera projec- tion model. Under this assumption, it can be shown that the set of 2D images produced by a 3D point feature of a rigid object can be optimally represented by two lines in the a ne space. Using this property, we reformulate the (human) face structure reconstruction problem in terms of the much familiar multiple baseline stereo matching problem [8]. Apart from the face modeling aspect, we also show how we use the results for reprojecting hu- man faces in identi cation tasks.

Original languageEnglish
Title of host publicationProceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998
PublisherIEEE Computer Society
Pages106-111
Number of pages6
ISBN (Print)0818683449, 9780818683442
DOIs
Publication statusPublished - 1998
Externally publishedYes
Event3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998 - Nara
Duration: 1998 Apr 141998 Apr 16

Other

Other3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998
CityNara
Period98/4/1498/4/16

Fingerprint

Cameras
Positive ions

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Sengupta, K., & Ohya, J. (1998). Human face structure estimation from multiple images using the 2D affine space. In Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998 (pp. 106-111). [670933] IEEE Computer Society. https://doi.org/10.1109/AFGR.1998.670933

Human face structure estimation from multiple images using the 2D affine space. / Sengupta, Kuntal; Ohya, Jun.

Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998. IEEE Computer Society, 1998. p. 106-111 670933.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sengupta, K & Ohya, J 1998, Human face structure estimation from multiple images using the 2D affine space. in Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998., 670933, IEEE Computer Society, pp. 106-111, 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998, Nara, 98/4/14. https://doi.org/10.1109/AFGR.1998.670933
Sengupta K, Ohya J. Human face structure estimation from multiple images using the 2D affine space. In Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998. IEEE Computer Society. 1998. p. 106-111. 670933 https://doi.org/10.1109/AFGR.1998.670933
Sengupta, Kuntal ; Ohya, Jun. / Human face structure estimation from multiple images using the 2D affine space. Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998. IEEE Computer Society, 1998. pp. 106-111
@inproceedings{e1926d87c4ec447c9a20cbe6501f56c0,
title = "Human face structure estimation from multiple images using the 2D affine space",
abstract = "In this paper we present an algorithm to estimate the human face structure. The input to the algorithm is not limited to an image sequence of a human head under rigid motion. It can be snapshots of the human face taken by the same or di erent cameras, over di erent periods of time. Since the depth variation of the human face is not very large, we use the a ne camera projec- tion model. Under this assumption, it can be shown that the set of 2D images produced by a 3D point feature of a rigid object can be optimally represented by two lines in the a ne space. Using this property, we reformulate the (human) face structure reconstruction problem in terms of the much familiar multiple baseline stereo matching problem [8]. Apart from the face modeling aspect, we also show how we use the results for reprojecting hu- man faces in identi cation tasks.",
author = "Kuntal Sengupta and Jun Ohya",
year = "1998",
doi = "10.1109/AFGR.1998.670933",
language = "English",
isbn = "0818683449",
pages = "106--111",
booktitle = "Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Human face structure estimation from multiple images using the 2D affine space

AU - Sengupta, Kuntal

AU - Ohya, Jun

PY - 1998

Y1 - 1998

N2 - In this paper we present an algorithm to estimate the human face structure. The input to the algorithm is not limited to an image sequence of a human head under rigid motion. It can be snapshots of the human face taken by the same or di erent cameras, over di erent periods of time. Since the depth variation of the human face is not very large, we use the a ne camera projec- tion model. Under this assumption, it can be shown that the set of 2D images produced by a 3D point feature of a rigid object can be optimally represented by two lines in the a ne space. Using this property, we reformulate the (human) face structure reconstruction problem in terms of the much familiar multiple baseline stereo matching problem [8]. Apart from the face modeling aspect, we also show how we use the results for reprojecting hu- man faces in identi cation tasks.

AB - In this paper we present an algorithm to estimate the human face structure. The input to the algorithm is not limited to an image sequence of a human head under rigid motion. It can be snapshots of the human face taken by the same or di erent cameras, over di erent periods of time. Since the depth variation of the human face is not very large, we use the a ne camera projec- tion model. Under this assumption, it can be shown that the set of 2D images produced by a 3D point feature of a rigid object can be optimally represented by two lines in the a ne space. Using this property, we reformulate the (human) face structure reconstruction problem in terms of the much familiar multiple baseline stereo matching problem [8]. Apart from the face modeling aspect, we also show how we use the results for reprojecting hu- man faces in identi cation tasks.

UR - http://www.scopus.com/inward/record.url?scp=11244283253&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=11244283253&partnerID=8YFLogxK

U2 - 10.1109/AFGR.1998.670933

DO - 10.1109/AFGR.1998.670933

M3 - Conference contribution

AN - SCOPUS:11244283253

SN - 0818683449

SN - 9780818683442

SP - 106

EP - 111

BT - Proceedings - 3rd IEEE International Conference on Automatic Face and Gesture Recognition, FG 1998

PB - IEEE Computer Society

ER -