Details
Originalsprache | Englisch |
---|---|
Titel des Sammelwerks | 30th AAAI Conference on Artificial Intelligence, AAAI 2016 |
Seiten | 3523-3529 |
Seitenumfang | 7 |
ISBN (elektronisch) | 9781577357605 |
Publikationsstatus | Veröffentlicht - 2016 |
Veranstaltung | 30th AAAI Conference on Artificial Intelligence, AAAI 2016 - Phoenix, USA / Vereinigte Staaten Dauer: 12 Feb. 2016 → 17 Feb. 2016 |
Publikationsreihe
Name | 30th AAAI Conference on Artificial Intelligence, AAAI 2016 |
---|
Abstract
Viewpoint estimation, especially in case of multiple object classes, remains an important and challenging problem. First, objects under different views undergo extreme appearance variations, often making within-class variance larger than between-class variance. Second, obtaining precise ground truth for real-world images, necessary for training supervised viewpoint estimation models, is extremely difficult and time consuming. As a result, annotated data is often available only for a limited number of classes. Hence it is desirable to share viewpoint information across classes. Additional complexity arises from unaligned pose labels between classes, i.e. a side view of a car might look more like a frontal view of a toaster, than its side view. To address these problems, we propose a metric learning approach for joint class prediction and pose estimation. Our approach allows to circumvent the problem of viewpoint alignment across multiple classes, and does not require dense viewpoint labels. Moreover, we show, that the learned metric generalizes to new classes, for which the pose labels are not available, and therefore makes it possible to use only partially annotated training sets, relying on the intrinsic similarities in the viewpoint manifolds. We evaluate our approach on two challenging multi-class datasets, 3DObjects and PASCAL3D+.
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Artificial intelligence
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
30th AAAI Conference on Artificial Intelligence, AAAI 2016. 2016. S. 3523-3529 (30th AAAI Conference on Artificial Intelligence, AAAI 2016).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Exploiting view-specific appearance similarities across classes for zero-shot pose prediction
T2 - 30th AAAI Conference on Artificial Intelligence, AAAI 2016
AU - Kuznetsova, Alina
AU - Hwang, Sung Ju
AU - Rosenhahn, Bodo
AU - Sigal, Leonid
PY - 2016
Y1 - 2016
N2 - Viewpoint estimation, especially in case of multiple object classes, remains an important and challenging problem. First, objects under different views undergo extreme appearance variations, often making within-class variance larger than between-class variance. Second, obtaining precise ground truth for real-world images, necessary for training supervised viewpoint estimation models, is extremely difficult and time consuming. As a result, annotated data is often available only for a limited number of classes. Hence it is desirable to share viewpoint information across classes. Additional complexity arises from unaligned pose labels between classes, i.e. a side view of a car might look more like a frontal view of a toaster, than its side view. To address these problems, we propose a metric learning approach for joint class prediction and pose estimation. Our approach allows to circumvent the problem of viewpoint alignment across multiple classes, and does not require dense viewpoint labels. Moreover, we show, that the learned metric generalizes to new classes, for which the pose labels are not available, and therefore makes it possible to use only partially annotated training sets, relying on the intrinsic similarities in the viewpoint manifolds. We evaluate our approach on two challenging multi-class datasets, 3DObjects and PASCAL3D+.
AB - Viewpoint estimation, especially in case of multiple object classes, remains an important and challenging problem. First, objects under different views undergo extreme appearance variations, often making within-class variance larger than between-class variance. Second, obtaining precise ground truth for real-world images, necessary for training supervised viewpoint estimation models, is extremely difficult and time consuming. As a result, annotated data is often available only for a limited number of classes. Hence it is desirable to share viewpoint information across classes. Additional complexity arises from unaligned pose labels between classes, i.e. a side view of a car might look more like a frontal view of a toaster, than its side view. To address these problems, we propose a metric learning approach for joint class prediction and pose estimation. Our approach allows to circumvent the problem of viewpoint alignment across multiple classes, and does not require dense viewpoint labels. Moreover, we show, that the learned metric generalizes to new classes, for which the pose labels are not available, and therefore makes it possible to use only partially annotated training sets, relying on the intrinsic similarities in the viewpoint manifolds. We evaluate our approach on two challenging multi-class datasets, 3DObjects and PASCAL3D+.
UR - http://www.scopus.com/inward/record.url?scp=84990043919&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84990043919
T3 - 30th AAAI Conference on Artificial Intelligence, AAAI 2016
SP - 3523
EP - 3529
BT - 30th AAAI Conference on Artificial Intelligence, AAAI 2016
Y2 - 12 February 2016 through 17 February 2016
ER -