Finding Near-Duplicate Videos in Large-Scale Collections
This chapter discusses the problem of Near-Duplicate Video Retrieval (NDVR). The main objective of a typical NDVR approach is: given a query video, retrieve all near-duplicate videos in a video repository and rank them based on their similarity to the query. Several approaches have been introduced i...
Main Authors: | , , , |
---|---|
Other Authors: | , , |
Format: | Book Part |
Language: | unknown |
Published: |
Springer
2019
|
Subjects: | |
Online Access: | https://qmro.qmul.ac.uk/xmlui/handle/123456789/61983 https://doi.org/10.1007/978-3-030-26752-0_4 |
id |
ftqueenmaryuniv:oai:qmro.qmul.ac.uk:123456789/61983 |
---|---|
record_format |
openpolar |
spelling |
ftqueenmaryuniv:oai:qmro.qmul.ac.uk:123456789/61983 2023-05-15T16:01:57+02:00 Finding Near-Duplicate Videos in Large-Scale Collections Kordopatis-Zilos, G Papadopoulos, S Patras, I Kompatsiaris, I Mezaris, V Nixon, L Papadopoulos, S Teyssou, D 2019-09-18 91 - 126 (45) https://qmro.qmul.ac.uk/xmlui/handle/123456789/61983 https://doi.org/10.1007/978-3-030-26752-0_4 unknown Springer Video Verification in the Fake News Era 978-3-030-26751-3 4 https://qmro.qmul.ac.uk/xmlui/handle/123456789/61983 doi:10.1007/978-3-030-26752-0_4 This is a pre-copyedited, author-produced version of an article accepted for publication in Video Verification in the Fake News Era following peer review. The version of record is available https://link.springer.com/chapter/10.1007%2F978-3-030-26752-0_4 © Springer Nature Switzerland AG 2019 Book chapter 2019 ftqueenmaryuniv https://doi.org/10.1007/978-3-030-26752-0_4 2022-09-25T20:19:04Z This chapter discusses the problem of Near-Duplicate Video Retrieval (NDVR). The main objective of a typical NDVR approach is: given a query video, retrieve all near-duplicate videos in a video repository and rank them based on their similarity to the query. Several approaches have been introduced in the literature, which can be roughly classified in three categories based on the level of video matching, i.e., (i) video-level, (ii) frame-level, and (iii) filter-and-refine matching. Two methods based on video-level matching are presented in this chapter. The first is an unsupervised scheme that relies on a modified Bag-of-Words (BoW) video representation. The second is a s upervised method based on Deep Metric Learning (DML). For the development of both methods, features are extracted from the intermediate layers of Convolutional Neural Networks and leveraged as frame descriptors, since they offer a compact and informative image representation, and lead to increased system efficiency. Extensive evaluation has been conducted on publicly available benchmark datasets, and the presented methods are compared with state-of-the-art approaches, achieving the best results in all evaluation setups. Book Part DML Queen Mary University of London: Queen Mary Research Online (QMRO) 91 126 Cham |
institution |
Open Polar |
collection |
Queen Mary University of London: Queen Mary Research Online (QMRO) |
op_collection_id |
ftqueenmaryuniv |
language |
unknown |
description |
This chapter discusses the problem of Near-Duplicate Video Retrieval (NDVR). The main objective of a typical NDVR approach is: given a query video, retrieve all near-duplicate videos in a video repository and rank them based on their similarity to the query. Several approaches have been introduced in the literature, which can be roughly classified in three categories based on the level of video matching, i.e., (i) video-level, (ii) frame-level, and (iii) filter-and-refine matching. Two methods based on video-level matching are presented in this chapter. The first is an unsupervised scheme that relies on a modified Bag-of-Words (BoW) video representation. The second is a s upervised method based on Deep Metric Learning (DML). For the development of both methods, features are extracted from the intermediate layers of Convolutional Neural Networks and leveraged as frame descriptors, since they offer a compact and informative image representation, and lead to increased system efficiency. Extensive evaluation has been conducted on publicly available benchmark datasets, and the presented methods are compared with state-of-the-art approaches, achieving the best results in all evaluation setups. |
author2 |
Mezaris, V Nixon, L Papadopoulos, S Teyssou, D |
format |
Book Part |
author |
Kordopatis-Zilos, G Papadopoulos, S Patras, I Kompatsiaris, I |
spellingShingle |
Kordopatis-Zilos, G Papadopoulos, S Patras, I Kompatsiaris, I Finding Near-Duplicate Videos in Large-Scale Collections |
author_facet |
Kordopatis-Zilos, G Papadopoulos, S Patras, I Kompatsiaris, I |
author_sort |
Kordopatis-Zilos, G |
title |
Finding Near-Duplicate Videos in Large-Scale Collections |
title_short |
Finding Near-Duplicate Videos in Large-Scale Collections |
title_full |
Finding Near-Duplicate Videos in Large-Scale Collections |
title_fullStr |
Finding Near-Duplicate Videos in Large-Scale Collections |
title_full_unstemmed |
Finding Near-Duplicate Videos in Large-Scale Collections |
title_sort |
finding near-duplicate videos in large-scale collections |
publisher |
Springer |
publishDate |
2019 |
url |
https://qmro.qmul.ac.uk/xmlui/handle/123456789/61983 https://doi.org/10.1007/978-3-030-26752-0_4 |
genre |
DML |
genre_facet |
DML |
op_relation |
Video Verification in the Fake News Era 978-3-030-26751-3 4 https://qmro.qmul.ac.uk/xmlui/handle/123456789/61983 doi:10.1007/978-3-030-26752-0_4 |
op_rights |
This is a pre-copyedited, author-produced version of an article accepted for publication in Video Verification in the Fake News Era following peer review. The version of record is available https://link.springer.com/chapter/10.1007%2F978-3-030-26752-0_4 © Springer Nature Switzerland AG 2019 |
op_doi |
https://doi.org/10.1007/978-3-030-26752-0_4 |
container_start_page |
91 |
op_container_end_page |
126 |
op_publisher_place |
Cham |
_version_ |
1766397612074729472 |