Skip to content

Semantically Selective Augmentation for Deep Compact Person Re-Identification

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Standard

Semantically Selective Augmentation for Deep Compact Person Re-Identification. / Ponce-López, Víctor; Burghardt, Tilo; Hannunna, Sion; Damen, Dima; Masullo, Alessandro; Mirmehdi, Majid.

Computer Vision – ECCV 2018 Workshops, Proceedings. ed. / Stefan Roth; Laura Leal-Taixé. Vol. 11130 Springer Verlag, 2019. p. 551-561 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11130 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Harvard

Ponce-López, V, Burghardt, T, Hannunna, S, Damen, D, Masullo, A & Mirmehdi, M 2019, Semantically Selective Augmentation for Deep Compact Person Re-Identification. in S Roth & L Leal-Taixé (eds), Computer Vision – ECCV 2018 Workshops, Proceedings. vol. 11130, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11130 LNCS, Springer Verlag, pp. 551-561, 15th European Conference on Computer Vision, ECCV 2018, Munich, Germany, 8/09/18. https://doi.org/10.1007/978-3-030-11012-3_41

APA

Ponce-López, V., Burghardt, T., Hannunna, S., Damen, D., Masullo, A., & Mirmehdi, M. (2019). Semantically Selective Augmentation for Deep Compact Person Re-Identification. In S. Roth, & L. Leal-Taixé (Eds.), Computer Vision – ECCV 2018 Workshops, Proceedings (Vol. 11130, pp. 551-561). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11130 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-11012-3_41

Vancouver

Ponce-López V, Burghardt T, Hannunna S, Damen D, Masullo A, Mirmehdi M. Semantically Selective Augmentation for Deep Compact Person Re-Identification. In Roth S, Leal-Taixé L, editors, Computer Vision – ECCV 2018 Workshops, Proceedings. Vol. 11130. Springer Verlag. 2019. p. 551-561. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-11012-3_41

Author

Ponce-López, Víctor ; Burghardt, Tilo ; Hannunna, Sion ; Damen, Dima ; Masullo, Alessandro ; Mirmehdi, Majid. / Semantically Selective Augmentation for Deep Compact Person Re-Identification. Computer Vision – ECCV 2018 Workshops, Proceedings. editor / Stefan Roth ; Laura Leal-Taixé. Vol. 11130 Springer Verlag, 2019. pp. 551-561 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

Bibtex

@inproceedings{8fffbb1fb5144898817a358a1fb2d0d1,
title = "Semantically Selective Augmentation for Deep Compact Person Re-Identification",
abstract = "We present a deep person re-identification approach that combines semantically selective, deep data augmentation with clustering-based network compression to generate high performance, light and fast inference networks. In particular, we propose to augment limited training data via sampling from a deep convolutional generative adversarial network (DCGAN), whose discriminator is constrained by a semantic classifier to explicitly control the domain specificity of the generation process. Thereby, we encode information in the classifier network which can be utilized to steer adversarial synthesis, and which fuels our CondenseNet ID-network training. We provide a quantitative and qualitative analysis of the approach and its variants on a number of datasets, obtaining results that outperform the state-of-the-art on the LIMA dataset for long-term monitoring in indoor living spaces.",
keywords = "Adversarial synthesis, Deep compression, Face filtering, Person re-identification, Selective augmentation",
author = "V{\'i}ctor Ponce-L{\'o}pez and Tilo Burghardt and Sion Hannunna and Dima Damen and Alessandro Masullo and Majid Mirmehdi",
year = "2019",
month = "1",
day = "29",
doi = "10.1007/978-3-030-11012-3_41",
language = "English",
isbn = "9783030110116",
volume = "11130",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "551--561",
editor = "Stefan Roth and Laura Leal-Taix{\'e}",
booktitle = "Computer Vision – ECCV 2018 Workshops, Proceedings",
address = "Germany",

}

RIS - suitable for import to EndNote

TY - GEN

T1 - Semantically Selective Augmentation for Deep Compact Person Re-Identification

AU - Ponce-López, Víctor

AU - Burghardt, Tilo

AU - Hannunna, Sion

AU - Damen, Dima

AU - Masullo, Alessandro

AU - Mirmehdi, Majid

PY - 2019/1/29

Y1 - 2019/1/29

N2 - We present a deep person re-identification approach that combines semantically selective, deep data augmentation with clustering-based network compression to generate high performance, light and fast inference networks. In particular, we propose to augment limited training data via sampling from a deep convolutional generative adversarial network (DCGAN), whose discriminator is constrained by a semantic classifier to explicitly control the domain specificity of the generation process. Thereby, we encode information in the classifier network which can be utilized to steer adversarial synthesis, and which fuels our CondenseNet ID-network training. We provide a quantitative and qualitative analysis of the approach and its variants on a number of datasets, obtaining results that outperform the state-of-the-art on the LIMA dataset for long-term monitoring in indoor living spaces.

AB - We present a deep person re-identification approach that combines semantically selective, deep data augmentation with clustering-based network compression to generate high performance, light and fast inference networks. In particular, we propose to augment limited training data via sampling from a deep convolutional generative adversarial network (DCGAN), whose discriminator is constrained by a semantic classifier to explicitly control the domain specificity of the generation process. Thereby, we encode information in the classifier network which can be utilized to steer adversarial synthesis, and which fuels our CondenseNet ID-network training. We provide a quantitative and qualitative analysis of the approach and its variants on a number of datasets, obtaining results that outperform the state-of-the-art on the LIMA dataset for long-term monitoring in indoor living spaces.

KW - Adversarial synthesis

KW - Deep compression

KW - Face filtering

KW - Person re-identification

KW - Selective augmentation

UR - http://www.scopus.com/inward/record.url?scp=85061827958&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-11012-3_41

DO - 10.1007/978-3-030-11012-3_41

M3 - Conference contribution

SN - 9783030110116

VL - 11130

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 551

EP - 561

BT - Computer Vision – ECCV 2018 Workshops, Proceedings

A2 - Roth, Stefan

A2 - Leal-Taixé, Laura

PB - Springer Verlag

ER -