Development of a Hardware Benchmark for Forensic Face Detection Applications

J Velasco-Mata; Deisy Chaves; V de Mata; M. W.  Al-Nabki; Eduardo Fidalgo; Enrique Alegre; George Azzopardi

doi:10.18239/jornadas_2021.34.28

Development of a Hardware Benchmark for Forensic Face Detection Applications

J Velasco-Mata, Deisy Chaves, V de Mata, M. W. Al-Nabki, Eduardo Fidalgo, Enrique Alegre, George Azzopardi

Information Systems

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

90 Downloads (Pure)

Abstract

Face detection techniques are valuable in the forensic investigation since they help criminal investigators to identify victims/offenders in child sexual exploitation material. Deep learning approaches proved successful in these tasks, but their
high computational requirements make them unsuitable if there are time constraints. To cope with this problem, we use a resizing strategy over three face detection techniques —MTCNN, PyramidBox and DSFD— to improve their speed over samples
selected from the WIDER Face and UFDD datasets across several CPUs and GPUs. The best speed-detection trade-off was achieved reducing the images to 50% of their original size and then applying DSFD. The fastest hardware for this purpose was a
Nvidia GPU based on the Turing architecture.

Original language	English
Title of host publication	Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales
Editors	Manuel A. Serrano
Publisher	Ediciones de la Universidad de Castilla-La Mancha
Pages	129-130
Number of pages	2
ISBN (Print)	978-84-9044-463-4
DOIs	https://doi.org/10.18239/jornadas_2021.34.28
Publication status	Published - Jun-2021
Event	Cybersecurity Research National Conferences - INCIBE, Leon, Spain Duration: 10-Jun-2021 → 11-Jun-2021

Publication series

Name	Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad
Volume	34

Conference

Conference	Cybersecurity Research National Conferences
Abbreviated title	JNIC
Country/Territory	Spain
City	Leon
Period	10/06/2021 → 11/06/2021

Access to Document

10.18239/jornadas_2021.34.28Licence: CC BY

Development of a Hardware Benchmark for Forensic Face Detection ApplicationsFinal publisher's version, 386 KBLicence: CC BY

Handle.net

Persistent link

Cite this

Velasco-Mata, J., Chaves, D., de Mata, V., Al-Nabki, M. W., Fidalgo, E., Alegre, E., & Azzopardi, G. (2021). Development of a Hardware Benchmark for Forensic Face Detection Applications. In M. A. Serrano (Ed.), Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales (pp. 129-130). (Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad; Vol. 34). Ediciones de la Universidad de Castilla-La Mancha. https://doi.org/10.18239/jornadas_2021.34.28

Velasco-Mata, J ; Chaves, Deisy ; de Mata, V et al. / Development of a Hardware Benchmark for Forensic Face Detection Applications. Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales. editor / Manuel A. Serrano. Ediciones de la Universidad de Castilla-La Mancha, 2021. pp. 129-130 (Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad).

@inproceedings{a6672e9a9c2b4c039cc29fabc42a173a,

title = "Development of a Hardware Benchmark for Forensic Face Detection Applications",

abstract = "Face detection techniques are valuable in the forensic investigation since they help criminal investigators to identify victims/offenders in child sexual exploitation material. Deep learning approaches proved successful in these tasks, but theirhigh computational requirements make them unsuitable if there are time constraints. To cope with this problem, we use a resizing strategy over three face detection techniques —MTCNN, PyramidBox and DSFD— to improve their speed over samplesselected from the WIDER Face and UFDD datasets across several CPUs and GPUs. The best speed-detection trade-off was achieved reducing the images to 50% of their original size and then applying DSFD. The fastest hardware for this purpose was aNvidia GPU based on the Turing architecture.",

author = "J Velasco-Mata and Deisy Chaves and {de Mata}, V and Al-Nabki, {M. W.} and Eduardo Fidalgo and Enrique Alegre and George Azzopardi",

year = "2021",

month = jun,

doi = "10.18239/jornadas_2021.34.28",

language = "English",

isbn = "978-84-9044-463-4",

series = "Investigaci{\'o}n en Ciberseguridad. Jornadas Nacionales de Investigaci{\'o}n en Ciberseguridad",

publisher = "Ediciones de la Universidad de Castilla-La Mancha",

pages = "129--130",

editor = "Serrano, {Manuel A.}",

booktitle = "Investigaci{\'o}n en Ciberseguridad Actas de las VI Jornadas Nacionales",

note = "Cybersecurity Research National Conferences, JNIC ; Conference date: 10-06-2021 Through 11-06-2021",

}

Velasco-Mata, J, Chaves, D, de Mata, V, Al-Nabki, MW, Fidalgo, E, Alegre, E & Azzopardi, G 2021, Development of a Hardware Benchmark for Forensic Face Detection Applications. in MA Serrano (ed.), Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales. Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad, vol. 34, Ediciones de la Universidad de Castilla-La Mancha, pp. 129-130, Cybersecurity Research National Conferences, Leon, Spain, 10/06/2021. https://doi.org/10.18239/jornadas_2021.34.28

Development of a Hardware Benchmark for Forensic Face Detection Applications. / Velasco-Mata, J; Chaves, Deisy; de Mata, V et al.
Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales. ed. / Manuel A. Serrano. Ediciones de la Universidad de Castilla-La Mancha, 2021. p. 129-130 (Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad; Vol. 34).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Development of a Hardware Benchmark for Forensic Face Detection Applications

AU - Velasco-Mata, J

AU - Chaves, Deisy

AU - de Mata, V

AU - Al-Nabki, M. W.

AU - Fidalgo, Eduardo

AU - Alegre, Enrique

AU - Azzopardi, George

PY - 2021/6

Y1 - 2021/6

N2 - Face detection techniques are valuable in the forensic investigation since they help criminal investigators to identify victims/offenders in child sexual exploitation material. Deep learning approaches proved successful in these tasks, but theirhigh computational requirements make them unsuitable if there are time constraints. To cope with this problem, we use a resizing strategy over three face detection techniques —MTCNN, PyramidBox and DSFD— to improve their speed over samplesselected from the WIDER Face and UFDD datasets across several CPUs and GPUs. The best speed-detection trade-off was achieved reducing the images to 50% of their original size and then applying DSFD. The fastest hardware for this purpose was aNvidia GPU based on the Turing architecture.

AB - Face detection techniques are valuable in the forensic investigation since they help criminal investigators to identify victims/offenders in child sexual exploitation material. Deep learning approaches proved successful in these tasks, but theirhigh computational requirements make them unsuitable if there are time constraints. To cope with this problem, we use a resizing strategy over three face detection techniques —MTCNN, PyramidBox and DSFD— to improve their speed over samplesselected from the WIDER Face and UFDD datasets across several CPUs and GPUs. The best speed-detection trade-off was achieved reducing the images to 50% of their original size and then applying DSFD. The fastest hardware for this purpose was aNvidia GPU based on the Turing architecture.

U2 - 10.18239/jornadas_2021.34.28

DO - 10.18239/jornadas_2021.34.28

M3 - Conference contribution

SN - 978-84-9044-463-4

T3 - Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad

SP - 129

EP - 130

BT - Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales

A2 - Serrano, Manuel A.

PB - Ediciones de la Universidad de Castilla-La Mancha

T2 - Cybersecurity Research National Conferences

Y2 - 10 June 2021 through 11 June 2021

ER -

Velasco-Mata J, Chaves D, de Mata V, Al-Nabki MW, Fidalgo E, Alegre E et al. Development of a Hardware Benchmark for Forensic Face Detection Applications. In Serrano MA, editor, Investigación en Ciberseguridad Actas de las VI Jornadas Nacionales. Ediciones de la Universidad de Castilla-La Mancha. 2021. p. 129-130. (Investigación en Ciberseguridad. Jornadas Nacionales de Investigación en Ciberseguridad). doi: 10.18239/jornadas_2021.34.28

Development of a Hardware Benchmark for Forensic Face Detection Applications

Abstract

Publication series

Conference

Access to Document

Handle.net

Fingerprint

Cite this