Near-Duplicate Video Retrieval with Deep Metric Learning
This work addresses the problem of Near-Duplicate Video Retrieval (NDVR). We propose an effective video-level NDVR scheme based on deep metric learning that leverages Convolutional Neural Network (CNN) features from intermediate layers to generate discriminative global video representations in tande...
Main Authors: | , , , |
---|---|
Format: | Conference Object |
Language: | unknown |
Published: |
2017
|
Subjects: | |
Online Access: | https://zenodo.org/record/852351 https://doi.org/10.5281/zenodo.852351 |
id |
ftzenodo:oai:zenodo.org:852351 |
---|---|
record_format |
openpolar |
spelling |
ftzenodo:oai:zenodo.org:852351 2023-05-15T16:01:55+02:00 Near-Duplicate Video Retrieval with Deep Metric Learning Kordopatis-Zilos, Giorgos Papadopoulos, Symeon Patras, Ioannis Kompatsiaris, Yiannis 2017-10-23 https://zenodo.org/record/852351 https://doi.org/10.5281/zenodo.852351 unknown info:eu-repo/grantAgreement/EC/H2020/687786/ doi:10.5281/zenodo.852350 https://zenodo.org/communities/invid-h2020 https://zenodo.org/record/852351 https://doi.org/10.5281/zenodo.852351 oai:zenodo.org:852351 info:eu-repo/semantics/openAccess https://creativecommons.org/licenses/by/4.0/legalcode Near-Duplicate Video Retrieval Deep Metric Learning info:eu-repo/semantics/conferencePaper publication-conferencepaper 2017 ftzenodo https://doi.org/10.5281/zenodo.85235110.5281/zenodo.852350 2023-03-11T00:41:01Z This work addresses the problem of Near-Duplicate Video Retrieval (NDVR). We propose an effective video-level NDVR scheme based on deep metric learning that leverages Convolutional Neural Network (CNN) features from intermediate layers to generate discriminative global video representations in tandem with a Deep Metric Learning (DML) framework with two fusion variations, trained to approximate an embedding function for accurate distance calculation between two near-duplicate videos. In contrast to most state-of-the-art methods, which exploit information deriving from the same source of data for both development and evaluation (which usually results to dataset-specific solutions), the proposed model is fed during training with sampled triplets generated from an independent dataset and is thoroughly tested on the widely used CC WEB VIDEO dataset, using two popular deep CNN architectures (AlexNet, GoogleNet). We demonstrate that the proposed approach achieves outstanding performance against the state-of-the-art, either with or without access to the evaluation dataset. Conference Object DML Zenodo Triplets ENVELOPE(-59.750,-59.750,-62.383,-62.383) |
institution |
Open Polar |
collection |
Zenodo |
op_collection_id |
ftzenodo |
language |
unknown |
topic |
Near-Duplicate Video Retrieval Deep Metric Learning |
spellingShingle |
Near-Duplicate Video Retrieval Deep Metric Learning Kordopatis-Zilos, Giorgos Papadopoulos, Symeon Patras, Ioannis Kompatsiaris, Yiannis Near-Duplicate Video Retrieval with Deep Metric Learning |
topic_facet |
Near-Duplicate Video Retrieval Deep Metric Learning |
description |
This work addresses the problem of Near-Duplicate Video Retrieval (NDVR). We propose an effective video-level NDVR scheme based on deep metric learning that leverages Convolutional Neural Network (CNN) features from intermediate layers to generate discriminative global video representations in tandem with a Deep Metric Learning (DML) framework with two fusion variations, trained to approximate an embedding function for accurate distance calculation between two near-duplicate videos. In contrast to most state-of-the-art methods, which exploit information deriving from the same source of data for both development and evaluation (which usually results to dataset-specific solutions), the proposed model is fed during training with sampled triplets generated from an independent dataset and is thoroughly tested on the widely used CC WEB VIDEO dataset, using two popular deep CNN architectures (AlexNet, GoogleNet). We demonstrate that the proposed approach achieves outstanding performance against the state-of-the-art, either with or without access to the evaluation dataset. |
format |
Conference Object |
author |
Kordopatis-Zilos, Giorgos Papadopoulos, Symeon Patras, Ioannis Kompatsiaris, Yiannis |
author_facet |
Kordopatis-Zilos, Giorgos Papadopoulos, Symeon Patras, Ioannis Kompatsiaris, Yiannis |
author_sort |
Kordopatis-Zilos, Giorgos |
title |
Near-Duplicate Video Retrieval with Deep Metric Learning |
title_short |
Near-Duplicate Video Retrieval with Deep Metric Learning |
title_full |
Near-Duplicate Video Retrieval with Deep Metric Learning |
title_fullStr |
Near-Duplicate Video Retrieval with Deep Metric Learning |
title_full_unstemmed |
Near-Duplicate Video Retrieval with Deep Metric Learning |
title_sort |
near-duplicate video retrieval with deep metric learning |
publishDate |
2017 |
url |
https://zenodo.org/record/852351 https://doi.org/10.5281/zenodo.852351 |
long_lat |
ENVELOPE(-59.750,-59.750,-62.383,-62.383) |
geographic |
Triplets |
geographic_facet |
Triplets |
genre |
DML |
genre_facet |
DML |
op_relation |
info:eu-repo/grantAgreement/EC/H2020/687786/ doi:10.5281/zenodo.852350 https://zenodo.org/communities/invid-h2020 https://zenodo.org/record/852351 https://doi.org/10.5281/zenodo.852351 oai:zenodo.org:852351 |
op_rights |
info:eu-repo/semantics/openAccess https://creativecommons.org/licenses/by/4.0/legalcode |
op_doi |
https://doi.org/10.5281/zenodo.85235110.5281/zenodo.852350 |
_version_ |
1766397594062290944 |