Learning sparse tag patterns for social image classification

User-generated tags associated with images from social media (e.g., Flickr) provide valuable textual resources for image classification. However, the noisy and huge tag vocabulary heavily degrades the effectiveness and efficiency of state-of-the-art image classification methods that exploited auxili...

Full description

Saved in:
Bibliographic Details
Main Authors: Duan, Ling-Yu, Yuan, Junsong, Li, Qingyong, Luo, Siwei, Lin, Jie
Other Authors: School of Electrical and Electronic Engineering
Format: Conference or Workshop Item
Language:English
Published: 2013
Subjects:
Online Access:https://hdl.handle.net/10356/79987
http://hdl.handle.net/10220/17899
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-79987
record_format dspace
spelling sg-ntu-dr.10356-799872020-03-07T13:24:43Z Learning sparse tag patterns for social image classification Duan, Ling-Yu Yuan, Junsong Li, Qingyong Luo, Siwei Lin, Jie School of Electrical and Electronic Engineering IEEE International Conference on Image Processing (19th : 2012 : Orlando, Florida, US) DRNTU::Engineering::Electrical and electronic engineering User-generated tags associated with images from social media (e.g., Flickr) provide valuable textual resources for image classification. However, the noisy and huge tag vocabulary heavily degrades the effectiveness and efficiency of state-of-the-art image classification methods that exploited auxiliary web data. To alleviate the problem, we introduce a Sparse Tag Patterns (STP) model to discover sparsity constrained co-occurrence tag patterns from large scale user contributed tags among social data. To fulfill the compactness and discriminability, we formulate STP as a problem of minimizing a quadratic loss function regularized by the bi-layer l1 norm. We treat the learned STP as alternative intermediate semantic image feature and verify its superiority within a search-based image classification framework. Experiments on 240K social images associated with millions of tags have demonstrated encouraging performance of the proposed method compared to the state-of-the-art. Accepted version 2013-11-29T03:31:26Z 2019-12-06T13:38:13Z 2013-11-29T03:31:26Z 2019-12-06T13:38:13Z 2012 2012 Conference Paper Lin, J., Duan, L.-Y., Yuan, J., Li, Q., & Luo, S. (2012). Learning sparse tag patterns for social image classification. 19th IEEE International Conference on Image Processing (ICIP), 2881-2884. https://hdl.handle.net/10356/79987 http://hdl.handle.net/10220/17899 10.1109/icip.2012.6467501 en © 2012 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: [http://dx.doi.org/10.1109/icip.2012.6467501]. 4 p. This work was supported in part by the National Basic Research Program of China (2009CB320902), in part by grants from the National Science Foundation of China (60902057 and 61121002) and Nanyang Assistant Professorship SUG M4080134. application/pdf
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic DRNTU::Engineering::Electrical and electronic engineering
spellingShingle DRNTU::Engineering::Electrical and electronic engineering
Duan, Ling-Yu
Yuan, Junsong
Li, Qingyong
Luo, Siwei
Lin, Jie
Learning sparse tag patterns for social image classification
description User-generated tags associated with images from social media (e.g., Flickr) provide valuable textual resources for image classification. However, the noisy and huge tag vocabulary heavily degrades the effectiveness and efficiency of state-of-the-art image classification methods that exploited auxiliary web data. To alleviate the problem, we introduce a Sparse Tag Patterns (STP) model to discover sparsity constrained co-occurrence tag patterns from large scale user contributed tags among social data. To fulfill the compactness and discriminability, we formulate STP as a problem of minimizing a quadratic loss function regularized by the bi-layer l1 norm. We treat the learned STP as alternative intermediate semantic image feature and verify its superiority within a search-based image classification framework. Experiments on 240K social images associated with millions of tags have demonstrated encouraging performance of the proposed method compared to the state-of-the-art.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Duan, Ling-Yu
Yuan, Junsong
Li, Qingyong
Luo, Siwei
Lin, Jie
format Conference or Workshop Item
author Duan, Ling-Yu
Yuan, Junsong
Li, Qingyong
Luo, Siwei
Lin, Jie
author_sort Duan, Ling-Yu
title Learning sparse tag patterns for social image classification
title_short Learning sparse tag patterns for social image classification
title_full Learning sparse tag patterns for social image classification
title_fullStr Learning sparse tag patterns for social image classification
title_full_unstemmed Learning sparse tag patterns for social image classification
title_sort learning sparse tag patterns for social image classification
publishDate 2013
url https://hdl.handle.net/10356/79987
http://hdl.handle.net/10220/17899
_version_ 1681041369875873792