Person De-identification in Activity Videos

M. Ivasic-Kos; A. Iosifidis; A. Tefas; Ioannis Pitas

doi:10.1109/MIPRO.2014.6859767

Person De-identification in Activity Videos

M. Ivasic-Kos
, A. Iosifidis
, A. Tefas
, Ioannis Pitas

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

12 Citations (Scopus)

354 Downloads (Pure)

Abstract

Person identification based on gait recognition has been extensively studied in the last two decades, while information appearing in different action types (like bend) has been recently exploited to this end. However, in most application scenarios it is sufficient to recognize the performed activity, whereas the ID of persons performing activities is not important. Since the same human body representations, e.g., body silhouettes, can be employed for both tasks, there is a need to automatically create privacy preserving representations. We have applied 2D Gaussian filtering to obfuscate the human body silhouettes that implicate information about the person ID. We have experimentally showed how the use of filtering affects the person identification and action recognition performance in different camera setups formed by an arbitrary number of cameras. In addition, the discriminative ability of different activities is examined and discussed in order to detect cases in which it is possible to apply Gaussian filter with a greater variance.

Original language	English
Title of host publication	2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014)
Subtitle of host publication	Proceedings of a meeting held 26-30 May 2014, Opatija, Croatia
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	1294-1299
Number of pages	6
ISBN (Electronic)	9789532330816
ISBN (Print)	9781479956579
DOIs	https://doi.org/10.1109/MIPRO.2014.6859767
Publication status	Published - Sept 2014
Event	IEEE International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO) - oPATIJA/aBBAZIA, Croatia Duration: 30 May 2014 → 3 Jun 2014

Conference

Conference	IEEE International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO)
Country/Territory	Croatia
City	oPATIJA/aBBAZIA
Period	30/05/14 → 3/06/14

Keywords

de-identification of human body silhouette
action recognition
Gaussian filter

Access to Document

10.1109/MIPRO.2014.6859767

Ioannis Pitas - Person De-identification in Activity Videos - 2014
This is the author accepted manuscript (AAM). The final published version (version of record) is available online via IEEE at 10.1109/MIPRO.2014.6859767. Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 632 KB

Handle.net

Persistent link

Cite this

Ivasic-Kos, M., Iosifidis, A., Tefas, A., & Pitas, I. (2014). Person De-identification in Activity Videos. In 2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014): Proceedings of a meeting held 26-30 May 2014, Opatija, Croatia (pp. 1294-1299 ). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/MIPRO.2014.6859767

@inproceedings{e59ac3d7239f49438c615313560242af,

title = "Person De-identification in Activity Videos",

abstract = "Person identification based on gait recognition has been extensively studied in the last two decades, while information appearing in different action types (like bend) has been recently exploited to this end. However, in most application scenarios it is sufficient to recognize the performed activity, whereas the ID of persons performing activities is not important. Since the same human body representations, e.g., body silhouettes, can be employed for both tasks, there is a need to automatically create privacy preserving representations. We have applied 2D Gaussian filtering to obfuscate the human body silhouettes that implicate information about the person ID. We have experimentally showed how the use of filtering affects the person identification and action recognition performance in different camera setups formed by an arbitrary number of cameras. In addition, the discriminative ability of different activities is examined and discussed in order to detect cases in which it is possible to apply Gaussian filter with a greater variance.",

keywords = "de-identification of human body silhouette, action recognition, Gaussian filter",

author = "M. Ivasic-Kos and A. Iosifidis and A. Tefas and Ioannis Pitas",

year = "2014",

month = sep,

doi = "10.1109/MIPRO.2014.6859767",

language = "English",

isbn = "9781479956579",

pages = "1294--1299 ",

booktitle = "2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

address = "United States",

note = "IEEE International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO) ; Conference date: 30-05-2014 Through 03-06-2014",

}

Ivasic-Kos, M, Iosifidis, A, Tefas, A & Pitas, I 2014, Person De-identification in Activity Videos. in 2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014): Proceedings of a meeting held 26-30 May 2014, Opatija, Croatia. Institute of Electrical and Electronics Engineers (IEEE), pp. 1294-1299 , IEEE International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), oPATIJA/aBBAZIA, Croatia, 30/05/14. https://doi.org/10.1109/MIPRO.2014.6859767

Person De-identification in Activity Videos. / Ivasic-Kos, M.; Iosifidis, A.; Tefas, A. et al.
2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014): Proceedings of a meeting held 26-30 May 2014, Opatija, Croatia. Institute of Electrical and Electronics Engineers (IEEE), 2014. p. 1294-1299 .

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Person De-identification in Activity Videos

AU - Ivasic-Kos, M.

AU - Iosifidis, A.

AU - Tefas, A.

AU - Pitas, Ioannis

PY - 2014/9

Y1 - 2014/9

N2 - Person identification based on gait recognition has been extensively studied in the last two decades, while information appearing in different action types (like bend) has been recently exploited to this end. However, in most application scenarios it is sufficient to recognize the performed activity, whereas the ID of persons performing activities is not important. Since the same human body representations, e.g., body silhouettes, can be employed for both tasks, there is a need to automatically create privacy preserving representations. We have applied 2D Gaussian filtering to obfuscate the human body silhouettes that implicate information about the person ID. We have experimentally showed how the use of filtering affects the person identification and action recognition performance in different camera setups formed by an arbitrary number of cameras. In addition, the discriminative ability of different activities is examined and discussed in order to detect cases in which it is possible to apply Gaussian filter with a greater variance.

AB - Person identification based on gait recognition has been extensively studied in the last two decades, while information appearing in different action types (like bend) has been recently exploited to this end. However, in most application scenarios it is sufficient to recognize the performed activity, whereas the ID of persons performing activities is not important. Since the same human body representations, e.g., body silhouettes, can be employed for both tasks, there is a need to automatically create privacy preserving representations. We have applied 2D Gaussian filtering to obfuscate the human body silhouettes that implicate information about the person ID. We have experimentally showed how the use of filtering affects the person identification and action recognition performance in different camera setups formed by an arbitrary number of cameras. In addition, the discriminative ability of different activities is examined and discussed in order to detect cases in which it is possible to apply Gaussian filter with a greater variance.

KW - de-identification of human body silhouette

KW - action recognition

KW - Gaussian filter

U2 - 10.1109/MIPRO.2014.6859767

DO - 10.1109/MIPRO.2014.6859767

M3 - Conference Contribution (Conference Proceeding)

SN - 9781479956579

SP - 1294

EP - 1299

BT - 2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014)

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - IEEE International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO)

Y2 - 30 May 2014 through 3 June 2014

ER -

Ivasic-Kos M, Iosifidis A, Tefas A, Pitas I. Person De-identification in Activity Videos. In 2014 37th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2014): Proceedings of a meeting held 26-30 May 2014, Opatija, Croatia. Institute of Electrical and Electronics Engineers (IEEE). 2014. p. 1294-1299 Epub 2014 Jul 24. doi: 10.1109/MIPRO.2014.6859767

Person De-identification in Activity Videos

Abstract

Conference

Keywords

Access to Document

Handle.net

Fingerprint

Cite this