TY - GEN
T1 - Exploring image specific structured loss for image annotation with incomplete labelling
AU - Xu, Xing
AU - Shimada, Atsushi
AU - Taniguch, Rin Ichiro
N1 - Funding Information:
This work was partly supported by Grant-in-Aid for Scientific Research (B), Grant Number 24300074. We thank reviewers for the precious comments.
Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - In this paper, we address the problem of image annotation with incomplete labelling, where the multiple objects in each training image are not fully labeled. The conventional one-versus-all SVM (OVA-SVM) that performs fairly well on full labelling decays drastically under the incomplete setting. Recently, structured learning method termed OVA-SSVM is proposed to boost the performance of OVA-SVM by modeling the structured associations of labels and show efficiency under incomplete setting. The OVA-SSVM assumes that each training sample includes a single label and adopts an loss measure of classification style that as long as one of the predicted label is correct, the overall prediction should be considered correct. However, this may not be appropriate for the multi-label annotation task. In this paper, we extend the OVA-SSVM method to the multi-label situation and design a novel image specific structured loss measure to account for the dependencies between predicted labels relying on the image-label associations. Then we develop an efficient optimization algorithm to learn the model parameters. Finally, we present extensive empirical results on two benchmark datasets with various degree of incompletion, and show that proposed method outperforms OVA-SSVM and achieves competitive performance compared with other state-of-the-art methods which are also designed for the issue of incomplete labelling.
AB - In this paper, we address the problem of image annotation with incomplete labelling, where the multiple objects in each training image are not fully labeled. The conventional one-versus-all SVM (OVA-SVM) that performs fairly well on full labelling decays drastically under the incomplete setting. Recently, structured learning method termed OVA-SSVM is proposed to boost the performance of OVA-SVM by modeling the structured associations of labels and show efficiency under incomplete setting. The OVA-SSVM assumes that each training sample includes a single label and adopts an loss measure of classification style that as long as one of the predicted label is correct, the overall prediction should be considered correct. However, this may not be appropriate for the multi-label annotation task. In this paper, we extend the OVA-SSVM method to the multi-label situation and design a novel image specific structured loss measure to account for the dependencies between predicted labels relying on the image-label associations. Then we develop an efficient optimization algorithm to learn the model parameters. Finally, we present extensive empirical results on two benchmark datasets with various degree of incompletion, and show that proposed method outperforms OVA-SSVM and achieves competitive performance compared with other state-of-the-art methods which are also designed for the issue of incomplete labelling.
UR - http://www.scopus.com/inward/record.url?scp=84938897005&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84938897005&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-16865-4_46
DO - 10.1007/978-3-319-16865-4_46
M3 - Conference contribution
AN - SCOPUS:84938897005
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 704
EP - 719
BT - Computer Vision - ACCV 2014 - 12th Asian Conference on Computer Vision, Revised Selected Papers
A2 - Reid, Ian
A2 - Yang, Ming-Hsuan
A2 - Saito, Hideo
A2 - Cremers, Daniel
PB - Springer Verlag
T2 - 12th Asian Conference on Computer Vision, ACCV 2014
Y2 - 1 November 2014 through 5 November 2014
ER -