DocumentCode :
1647481
Title :
Multi-feature Late Fusion for Image Tagging
Author :
Xi Liu ; Rujie Liu ; Qiong Cao ; Fei Li
Author_Institution :
Fujitsu R&D Center Co., Ltd., Beijing, China
fYear :
2013
Firstpage :
34
Lastpage :
37
Abstract :
Image tagging plays a critical role in image indexing and retrieval and it has gained more and more attention along with the increasing availability of large quantities of web images. However, most of current tagging methods only utilize single feature type, while combining multiple types of features has been proved to be effective for image analysis. In this paper, we propose a multi-feature late fusion method for image tagging. For an image, we first learn several scores with regard to each tag by using different single features or combinations of single features based on a tag relevance learner. Then we learn an optimal combination weight for each tag score and linearly combine all the tag scores with the learned weights. Finally, a low-rank tag pair wise matrix is learned with the linearly combined tag scores and a robust tag score is recovered from the low-rank matrix. The tags with the largest scores are regarded as the predicted tags. We compare our approach with several multi-feature fusion techniques over a real-world dataset NUSWIDE and show the effectiveness of the proposed multi-feature fusion method.
Keywords :
image fusion; image retrieval; indexing; learning (artificial intelligence); matrix algebra; relevance feedback; NUSWIDE dataset; Web images; image analysis; image indexing; image retrieval; image tagging; low-rank tag pairwise matrix; multifeature late fusion; multiple feature types; single feature type; tag relevance learner; tag score; Feature extraction; Image color analysis; Robustness; Sparse matrices; Tagging; Training; Visualization; Image tagging; Late fusion; Low rank; Multi-feature; Tag relevance;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition (ACPR), 2013 2nd IAPR Asian Conference on
Conference_Location :
Naha
Type :
conf
DOI :
10.1109/ACPR.2013.25
Filename :
6778277
Link To Document :
بازگشت