Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition
As an important research direction in image and video processing, set-based video recognition requires speed and accuracy. However, the existing static modeling methods focus on computational speed but ignore accuracy, whereas the dynamic modeling methods are higher-accuracy but ignore the computati...
Published in: | Applied Sciences |
---|---|
Main Authors: | , , , , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3390/app13116383 |
id |
ftmdpi:oai:mdpi.com:/2076-3417/13/11/6383/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2076-3417/13/11/6383/ 2023-08-20T04:06:10+02:00 Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition Xizhan Gao Kang Wei Jia Li Ziyu Shi Hui Zhao Sijie Niu agris 2023-05-23 application/pdf https://doi.org/10.3390/app13116383 EN eng Multidisciplinary Digital Publishing Institute Computing and Artificial Intelligence https://dx.doi.org/10.3390/app13116383 https://creativecommons.org/licenses/by/4.0/ Applied Sciences; Volume 13; Issue 11; Pages: 6383 set-based video recognition image set classification manifold learning fast and accurate classification discriminative dictionary learning Text 2023 ftmdpi https://doi.org/10.3390/app13116383 2023-08-01T10:11:26Z As an important research direction in image and video processing, set-based video recognition requires speed and accuracy. However, the existing static modeling methods focus on computational speed but ignore accuracy, whereas the dynamic modeling methods are higher-accuracy but ignore the computational speed. Combining these two types of methods to obtain fast and accurate recognition results remains a challenging problem. Motivated by this, in this study, a novel Manifolds-based Low-Rank Dictionary Pair Learning (MbLRDPL) method was developed for a set-based video recognition/image set classification task. Specifically, each video or image set was first modeled as a covariance matrix or linear subspace, which can be seen as a point on a Riemannian manifold. Second, the proposed MbLRDPL learned discriminative class-specific synthesis and analysis dictionaries by clearly imposing the nuclear norm on the synthesis dictionaries. The experimental results show that our method achieved the best classification accuracy (100%, 72.16%, 95%) on three datasets with the fastest computing time, reducing the errors of state-of-the-art methods (JMLC, DML, CEBSR) by 0.96–75.69%. Text DML MDPI Open Access Publishing Applied Sciences 13 11 6383 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
set-based video recognition image set classification manifold learning fast and accurate classification discriminative dictionary learning |
spellingShingle |
set-based video recognition image set classification manifold learning fast and accurate classification discriminative dictionary learning Xizhan Gao Kang Wei Jia Li Ziyu Shi Hui Zhao Sijie Niu Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
topic_facet |
set-based video recognition image set classification manifold learning fast and accurate classification discriminative dictionary learning |
description |
As an important research direction in image and video processing, set-based video recognition requires speed and accuracy. However, the existing static modeling methods focus on computational speed but ignore accuracy, whereas the dynamic modeling methods are higher-accuracy but ignore the computational speed. Combining these two types of methods to obtain fast and accurate recognition results remains a challenging problem. Motivated by this, in this study, a novel Manifolds-based Low-Rank Dictionary Pair Learning (MbLRDPL) method was developed for a set-based video recognition/image set classification task. Specifically, each video or image set was first modeled as a covariance matrix or linear subspace, which can be seen as a point on a Riemannian manifold. Second, the proposed MbLRDPL learned discriminative class-specific synthesis and analysis dictionaries by clearly imposing the nuclear norm on the synthesis dictionaries. The experimental results show that our method achieved the best classification accuracy (100%, 72.16%, 95%) on three datasets with the fastest computing time, reducing the errors of state-of-the-art methods (JMLC, DML, CEBSR) by 0.96–75.69%. |
format |
Text |
author |
Xizhan Gao Kang Wei Jia Li Ziyu Shi Hui Zhao Sijie Niu |
author_facet |
Xizhan Gao Kang Wei Jia Li Ziyu Shi Hui Zhao Sijie Niu |
author_sort |
Xizhan Gao |
title |
Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
title_short |
Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
title_full |
Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
title_fullStr |
Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
title_full_unstemmed |
Manifolds-Based Low-Rank Dictionary Pair Learning for Efficient Set-Based Video Recognition |
title_sort |
manifolds-based low-rank dictionary pair learning for efficient set-based video recognition |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2023 |
url |
https://doi.org/10.3390/app13116383 |
op_coverage |
agris |
genre |
DML |
genre_facet |
DML |
op_source |
Applied Sciences; Volume 13; Issue 11; Pages: 6383 |
op_relation |
Computing and Artificial Intelligence https://dx.doi.org/10.3390/app13116383 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/app13116383 |
container_title |
Applied Sciences |
container_volume |
13 |
container_issue |
11 |
container_start_page |
6383 |
_version_ |
1774717097777037312 |