Towards understanding the character of quality sampling in deep learning face recognition
Abstract Face recognition has become one of the most important modalities of biometrics in recent years. It widely utilises deep learning computer vision tools and adopts large collections of unconstrained face images of celebrities for training. Such choice of the data is related to its public avai...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-09-01
|
| Series: | IET Biometrics |
| Online Access: | https://doi.org/10.1049/bme2.12095 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832559645033496576 |
|---|---|
| author | Iurii Medvedev João Tremoço Beatriz Mano Luís Espírito Santo Nuno Gonçalves |
| author_facet | Iurii Medvedev João Tremoço Beatriz Mano Luís Espírito Santo Nuno Gonçalves |
| author_sort | Iurii Medvedev |
| collection | DOAJ |
| description | Abstract Face recognition has become one of the most important modalities of biometrics in recent years. It widely utilises deep learning computer vision tools and adopts large collections of unconstrained face images of celebrities for training. Such choice of the data is related to its public availability when existing document compliant face image collections are hardly accessible due to security and privacy issues. Such inconsistency between the training data and deploy scenario may lead to a leak in performance in biometric systems, which are developed specifically for dealing with ID document compliant images. To mitigate this problem, we propose to regularise the training of the deep face recognition network with a specific sample mining strategy, which penalises the samples by their estimated quality. In addition to several considered quality metrics in recent work, we also expand our deep learning strategy to other sophisticated quality estimation methods and perform experiments to better understand the nature of quality sampling. Namely, we seek for the penalising manner (sampling character) that better satisfies the purpose of adapting deep learning face recognition for images of ID and travel documents. Extensive experiments demonstrate the efficiency of the approach for ID document compliant face images. |
| format | Article |
| id | doaj-art-ac8723a52b9e42f2b023b661d52e6ef4 |
| institution | Kabale University |
| issn | 2047-4938 2047-4946 |
| language | English |
| publishDate | 2022-09-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Biometrics |
| spelling | doaj-art-ac8723a52b9e42f2b023b661d52e6ef42025-02-03T01:29:39ZengWileyIET Biometrics2047-49382047-49462022-09-0111549851110.1049/bme2.12095Towards understanding the character of quality sampling in deep learning face recognitionIurii Medvedev0João Tremoço1Beatriz Mano2Luís Espírito Santo3Nuno Gonçalves4Institute of Systems and Robotics University of Coimbra Coimbra PortugalInstitute of Systems and Robotics University of Coimbra Coimbra PortugalPortuguese Mint and Official Printing Office (INCM) Lisbon PortugalPortuguese Mint and Official Printing Office (INCM) Lisbon PortugalInstitute of Systems and Robotics University of Coimbra Coimbra PortugalAbstract Face recognition has become one of the most important modalities of biometrics in recent years. It widely utilises deep learning computer vision tools and adopts large collections of unconstrained face images of celebrities for training. Such choice of the data is related to its public availability when existing document compliant face image collections are hardly accessible due to security and privacy issues. Such inconsistency between the training data and deploy scenario may lead to a leak in performance in biometric systems, which are developed specifically for dealing with ID document compliant images. To mitigate this problem, we propose to regularise the training of the deep face recognition network with a specific sample mining strategy, which penalises the samples by their estimated quality. In addition to several considered quality metrics in recent work, we also expand our deep learning strategy to other sophisticated quality estimation methods and perform experiments to better understand the nature of quality sampling. Namely, we seek for the penalising manner (sampling character) that better satisfies the purpose of adapting deep learning face recognition for images of ID and travel documents. Extensive experiments demonstrate the efficiency of the approach for ID document compliant face images.https://doi.org/10.1049/bme2.12095 |
| spellingShingle | Iurii Medvedev João Tremoço Beatriz Mano Luís Espírito Santo Nuno Gonçalves Towards understanding the character of quality sampling in deep learning face recognition IET Biometrics |
| title | Towards understanding the character of quality sampling in deep learning face recognition |
| title_full | Towards understanding the character of quality sampling in deep learning face recognition |
| title_fullStr | Towards understanding the character of quality sampling in deep learning face recognition |
| title_full_unstemmed | Towards understanding the character of quality sampling in deep learning face recognition |
| title_short | Towards understanding the character of quality sampling in deep learning face recognition |
| title_sort | towards understanding the character of quality sampling in deep learning face recognition |
| url | https://doi.org/10.1049/bme2.12095 |
| work_keys_str_mv | AT iuriimedvedev towardsunderstandingthecharacterofqualitysamplingindeeplearningfacerecognition AT joaotremoco towardsunderstandingthecharacterofqualitysamplingindeeplearningfacerecognition AT beatrizmano towardsunderstandingthecharacterofqualitysamplingindeeplearningfacerecognition AT luisespiritosanto towardsunderstandingthecharacterofqualitysamplingindeeplearningfacerecognition AT nunogoncalves towardsunderstandingthecharacterofqualitysamplingindeeplearningfacerecognition |