TY - CHAP

T1 - Efficient Classification of Images with Taxonomies

AU - Binder, Alexander

AU - Kawanabe, Motoaki

AU - Brefeld, Ulf

N1 - Conference code: 9

PY - 2010

Y1 - 2010

N2 - We study the problem of classifying images into a given, pre-determined taxonomy. The task can be elegantly translated into the structured learning framework. Structured learning, however, is known for its memory consuming and slow training processes. The contribution of our paper is twofold: Firstly, we propose an e.cient decomposition of the structured learning approach into an equivalent ensemble of local support vector machines (SVMs) which can be trained with standard techniques. Secondly, we combine the local SVMs to a global model by re-incorporating the taxonomy into the training process. Our empirical results on Caltech256 and VOC2006 data show that our local-global SVM effectively exploits the structure of the taxonomy and outperforms multi-class classification approaches.

AB - We study the problem of classifying images into a given, pre-determined taxonomy. The task can be elegantly translated into the structured learning framework. Structured learning, however, is known for its memory consuming and slow training processes. The contribution of our paper is twofold: Firstly, we propose an e.cient decomposition of the structured learning approach into an equivalent ensemble of local support vector machines (SVMs) which can be trained with standard techniques. Secondly, we combine the local SVMs to a global model by re-incorporating the taxonomy into the training process. Our empirical results on Caltech256 and VOC2006 data show that our local-global SVM effectively exploits the structure of the taxonomy and outperforms multi-class classification approaches.

KW - Informatics

KW - Empirical results

KW - Global models

KW - Local support

KW - Local-global

KW - Multi-class classification

KW - Structured learning

KW - Training process

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=78650463536&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-12297-2_34

DO - 10.1007/978-3-642-12297-2_34

M3 - Article in conference proceedings

SN - 978-3-642-12296-5

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 351

EP - 362

BT - Computer Vision, ACCV 2009 - 9th Asian Conference on Computer Vision, Revised Selected Papers

PB - Springer

CY - Berlin, Heidelberg

T2 - 9th Asian Conference on Computer Vision - ACCV 2009

Y2 - 23 September 2009 through 27 September 2009

ER -