A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning
Abstract Video data are usually represented by high dimensional features. The performance of video semantic recognition, however, may be deteriorated due to the irrelevant and redundant components included into the high dimensional representations. To improve the performance of video semantic recogn...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
SpringerOpen
2019-02-01
|
Series: | EURASIP Journal on Image and Video Processing |
Subjects: | |
Online Access: | http://link.springer.com/article/10.1186/s13640-019-0428-5 |
id |
doaj-48ff732699624f66a73c10d78d8efef2 |
---|---|
record_format |
Article |
spelling |
doaj-48ff732699624f66a73c10d78d8efef22020-11-25T02:56:53ZengSpringerOpenEURASIP Journal on Image and Video Processing1687-52812019-02-012019111510.1186/s13640-019-0428-5A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learningJianguang Zhang0Yahong Han1Jianmin Jiang2Zhongrun Zhou3Da An4JieJing Liu5Zhifei Song6College of Mathematics and Computer Science, Hengshui UniversityCollege of Intelligence and Computing, Tianjin UniversityCollege of Computer Science & Software Engineering, Shenzhen UniversityHong Kong Applied Science and Technology Research InstituteCollege of Mathematics and Computer Science, Hengshui UniversityOffice of Academic Affaires, Hengshui UniversityOffice of Academic Research, Hengshui UniversityAbstract Video data are usually represented by high dimensional features. The performance of video semantic recognition, however, may be deteriorated due to the irrelevant and redundant components included into the high dimensional representations. To improve the performance of video semantic recognition, we propose a new feature selection framework in this paper and validate it through applications of video semantic recognition. Two issues are considered in our framework. First, while those labeled videos are precious, their relevant labeled images are abundant and available in the WEB. Therefore, a supervised transfer learning is proposed to achieve the cross-media analysis, in which the discriminative features are selected by evaluating feature’s correlation with the classes of videos and relevant images. Second, the labeled videos are normally rare in real-world applications. In our framework, therefore, an unsupervised subspace learning is added to retain the most valuable information and eliminate the feature redundancies by leveraging both labeled and unlabeled videos. The cross-media analysis and embedded learning are simultaneously learned in a joint framework, which enables our algorithm to utilize the common knowledge of cross-media analysis and embedded learning as supplementary information to facilitate decision making. An efficient iterative algorithm is proposed to optimize the proposed learning-based feature selection, in which convergence is guaranteed. Experiments on different databases have demonstrated the effectiveness of the proposed algorithm.http://link.springer.com/article/10.1186/s13640-019-0428-5Feature selectionCross-media analysisEmbedded learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Jianguang Zhang Yahong Han Jianmin Jiang Zhongrun Zhou Da An JieJing Liu Zhifei Song |
spellingShingle |
Jianguang Zhang Yahong Han Jianmin Jiang Zhongrun Zhou Da An JieJing Liu Zhifei Song A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning EURASIP Journal on Image and Video Processing Feature selection Cross-media analysis Embedded learning |
author_facet |
Jianguang Zhang Yahong Han Jianmin Jiang Zhongrun Zhou Da An JieJing Liu Zhifei Song |
author_sort |
Jianguang Zhang |
title |
A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
title_short |
A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
title_full |
A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
title_fullStr |
A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
title_full_unstemmed |
A feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
title_sort |
feature selection framework for video semantic recognition via integrated cross-media analysis and embedded learning |
publisher |
SpringerOpen |
series |
EURASIP Journal on Image and Video Processing |
issn |
1687-5281 |
publishDate |
2019-02-01 |
description |
Abstract Video data are usually represented by high dimensional features. The performance of video semantic recognition, however, may be deteriorated due to the irrelevant and redundant components included into the high dimensional representations. To improve the performance of video semantic recognition, we propose a new feature selection framework in this paper and validate it through applications of video semantic recognition. Two issues are considered in our framework. First, while those labeled videos are precious, their relevant labeled images are abundant and available in the WEB. Therefore, a supervised transfer learning is proposed to achieve the cross-media analysis, in which the discriminative features are selected by evaluating feature’s correlation with the classes of videos and relevant images. Second, the labeled videos are normally rare in real-world applications. In our framework, therefore, an unsupervised subspace learning is added to retain the most valuable information and eliminate the feature redundancies by leveraging both labeled and unlabeled videos. The cross-media analysis and embedded learning are simultaneously learned in a joint framework, which enables our algorithm to utilize the common knowledge of cross-media analysis and embedded learning as supplementary information to facilitate decision making. An efficient iterative algorithm is proposed to optimize the proposed learning-based feature selection, in which convergence is guaranteed. Experiments on different databases have demonstrated the effectiveness of the proposed algorithm. |
topic |
Feature selection Cross-media analysis Embedded learning |
url |
http://link.springer.com/article/10.1186/s13640-019-0428-5 |
work_keys_str_mv |
AT jianguangzhang afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT yahonghan afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT jianminjiang afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT zhongrunzhou afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT daan afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT jiejingliu afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT zhifeisong afeatureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT jianguangzhang featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT yahonghan featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT jianminjiang featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT zhongrunzhou featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT daan featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT jiejingliu featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning AT zhifeisong featureselectionframeworkforvideosemanticrecognitionviaintegratedcrossmediaanalysisandembeddedlearning |
_version_ |
1724711805124608000 |