Abstract
We consider the problem of object recognition in 3D using an ensemble of attribute-based classifiers. We propose two new concepts to improve classification in practical situations, and show their implementation in an approach implemented for recognition from point-cloud data. First, the viewing conditions can have a strong influence on classification performance. We study the impact of the distance between the camera and the object and propose an approach to fusing multiple attribute classifiers, which incorporates distance into the decision making. Second, lack of representative training samples often makes it difficult to learn the optimal threshold value for best positive and negative detection rate. We address this issue, by setting in our attribute classifiers instead of just one threshold value, two threshold values to distinguish a positive, a negative and an uncertainty class, and we prove the theoretical correctness of this approach. Empirical studies demonstrate the effectiveness and feasibility of the proposed concepts.
| Original language | English (US) |
|---|---|
| Title of host publication | Computer Vision - 14th European Conference, ECCV 2016, Proceedings |
| Publisher | Springer Verlag |
| Pages | 801-815 |
| Number of pages | 15 |
| Volume | 9907 LNCS |
| ISBN (Print) | 9783319464862 |
| DOIs | |
| State | Published - 2016 |
| Externally published | Yes |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 9907 LNCS |
| ISSN (Print) | 03029743 |
| ISSN (Electronic) | 16113349 |
Fingerprint
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science(all)
Cite this
Reliable attribute-based object recognition using high predictive value classifiers. / Luan, Wentao; Yang, Yezhou; Fermüller, Cornelia; Baras, John S.
Computer Vision - 14th European Conference, ECCV 2016, Proceedings. Vol. 9907 LNCS Springer Verlag, 2016. p. 801-815 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9907 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Reliable attribute-based object recognition using high predictive value classifiers
AU - Luan, Wentao
AU - Yang, Yezhou
AU - Fermüller, Cornelia
AU - Baras, John S.
PY - 2016
Y1 - 2016
N2 - We consider the problem of object recognition in 3D using an ensemble of attribute-based classifiers. We propose two new concepts to improve classification in practical situations, and show their implementation in an approach implemented for recognition from point-cloud data. First, the viewing conditions can have a strong influence on classification performance. We study the impact of the distance between the camera and the object and propose an approach to fusing multiple attribute classifiers, which incorporates distance into the decision making. Second, lack of representative training samples often makes it difficult to learn the optimal threshold value for best positive and negative detection rate. We address this issue, by setting in our attribute classifiers instead of just one threshold value, two threshold values to distinguish a positive, a negative and an uncertainty class, and we prove the theoretical correctness of this approach. Empirical studies demonstrate the effectiveness and feasibility of the proposed concepts.
AB - We consider the problem of object recognition in 3D using an ensemble of attribute-based classifiers. We propose two new concepts to improve classification in practical situations, and show their implementation in an approach implemented for recognition from point-cloud data. First, the viewing conditions can have a strong influence on classification performance. We study the impact of the distance between the camera and the object and propose an approach to fusing multiple attribute classifiers, which incorporates distance into the decision making. Second, lack of representative training samples often makes it difficult to learn the optimal threshold value for best positive and negative detection rate. We address this issue, by setting in our attribute classifiers instead of just one threshold value, two threshold values to distinguish a positive, a negative and an uncertainty class, and we prove the theoretical correctness of this approach. Empirical studies demonstrate the effectiveness and feasibility of the proposed concepts.
UR - http://www.scopus.com/inward/record.url?scp=84990041919&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84990041919&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-46487-9_49
DO - 10.1007/978-3-319-46487-9_49
M3 - Conference contribution
AN - SCOPUS:84990041919
SN - 9783319464862
VL - 9907 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 801
EP - 815
BT - Computer Vision - 14th European Conference, ECCV 2016, Proceedings
PB - Springer Verlag
ER -