Dual Minkowski Loss for Face Verification of Convolutional Network
Despite face recognition and verification have achieved great success in recent years, these achievements are experimental results on fixed data sets. Implementing these outstanding technologies in the field of undeveloped data sets presents serious challenges. We adopt three state-of-the-art pre-tr...
Published in: | MATEC Web of Conferences |
---|---|
Main Authors: | , |
Format: | Article in Journal/Newspaper |
Language: | English French |
Published: |
EDP Sciences
2018
|
Subjects: | |
Online Access: | https://doi.org/10.1051/matecconf/201823201007 https://doaj.org/article/a959fb0834054b89aa9d7ab8a5803f63 |
id |
ftdoajarticles:oai:doaj.org/article:a959fb0834054b89aa9d7ab8a5803f63 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:a959fb0834054b89aa9d7ab8a5803f63 2023-05-15T16:01:52+02:00 Dual Minkowski Loss for Face Verification of Convolutional Network Wang Dandan Chen Yan 2018-01-01T00:00:00Z https://doi.org/10.1051/matecconf/201823201007 https://doaj.org/article/a959fb0834054b89aa9d7ab8a5803f63 EN FR eng fre EDP Sciences https://doi.org/10.1051/matecconf/201823201007 https://doaj.org/toc/2261-236X 2261-236X doi:10.1051/matecconf/201823201007 https://doaj.org/article/a959fb0834054b89aa9d7ab8a5803f63 MATEC Web of Conferences, Vol 232, p 01007 (2018) Engineering (General). Civil engineering (General) TA1-2040 article 2018 ftdoajarticles https://doi.org/10.1051/matecconf/201823201007 2022-12-31T13:51:21Z Despite face recognition and verification have achieved great success in recent years, these achievements are experimental results on fixed data sets. Implementing these outstanding technologies in the field of undeveloped data sets presents serious challenges. We adopt three state-of-the-art pre-trained models on an entire new dataset University Test System Database (UTSD), however the results are far from satisfactory. Therefore, two methods are adopted to solve this problem. The first way is data augmentation including horizontal flipping, cropping and RGB channels transform, which can solve the imbalance of label pairs. The second way is the combination of Manhattan Distance and Euclidean Distance, we call it Dual Minkowski Loss (DML). Through the implementation of photo augmentation and innovative method on UTSD, the accuracy of face verification has been significantly improved, achieving the best 99.3%. Article in Journal/Newspaper DML Directory of Open Access Journals: DOAJ Articles MATEC Web of Conferences 232 01007 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English French |
topic |
Engineering (General). Civil engineering (General) TA1-2040 |
spellingShingle |
Engineering (General). Civil engineering (General) TA1-2040 Wang Dandan Chen Yan Dual Minkowski Loss for Face Verification of Convolutional Network |
topic_facet |
Engineering (General). Civil engineering (General) TA1-2040 |
description |
Despite face recognition and verification have achieved great success in recent years, these achievements are experimental results on fixed data sets. Implementing these outstanding technologies in the field of undeveloped data sets presents serious challenges. We adopt three state-of-the-art pre-trained models on an entire new dataset University Test System Database (UTSD), however the results are far from satisfactory. Therefore, two methods are adopted to solve this problem. The first way is data augmentation including horizontal flipping, cropping and RGB channels transform, which can solve the imbalance of label pairs. The second way is the combination of Manhattan Distance and Euclidean Distance, we call it Dual Minkowski Loss (DML). Through the implementation of photo augmentation and innovative method on UTSD, the accuracy of face verification has been significantly improved, achieving the best 99.3%. |
format |
Article in Journal/Newspaper |
author |
Wang Dandan Chen Yan |
author_facet |
Wang Dandan Chen Yan |
author_sort |
Wang Dandan |
title |
Dual Minkowski Loss for Face Verification of Convolutional Network |
title_short |
Dual Minkowski Loss for Face Verification of Convolutional Network |
title_full |
Dual Minkowski Loss for Face Verification of Convolutional Network |
title_fullStr |
Dual Minkowski Loss for Face Verification of Convolutional Network |
title_full_unstemmed |
Dual Minkowski Loss for Face Verification of Convolutional Network |
title_sort |
dual minkowski loss for face verification of convolutional network |
publisher |
EDP Sciences |
publishDate |
2018 |
url |
https://doi.org/10.1051/matecconf/201823201007 https://doaj.org/article/a959fb0834054b89aa9d7ab8a5803f63 |
genre |
DML |
genre_facet |
DML |
op_source |
MATEC Web of Conferences, Vol 232, p 01007 (2018) |
op_relation |
https://doi.org/10.1051/matecconf/201823201007 https://doaj.org/toc/2261-236X 2261-236X doi:10.1051/matecconf/201823201007 https://doaj.org/article/a959fb0834054b89aa9d7ab8a5803f63 |
op_doi |
https://doi.org/10.1051/matecconf/201823201007 |
container_title |
MATEC Web of Conferences |
container_volume |
232 |
container_start_page |
01007 |
_version_ |
1766397569525612544 |