Skip to content

Semantically Selective Augmentation for Deep Compact Person Re-Identification

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

Original languageEnglish
Title of host publicationComputer Vision – ECCV 2018 Workshops, Proceedings
EditorsStefan Roth, Laura Leal-Taixé
Publisher or commissioning bodySpringer Verlag
Pages551-561
Number of pages11
Volume11130
ISBN (Electronic)978-3-030-11012-3
ISBN (Print)9783030110116
DOIs
DateAccepted/In press - 25 Jul 2018
DatePublished (current) - 29 Jan 2019
Event15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany
Duration: 8 Sep 201814 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11130 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th European Conference on Computer Vision, ECCV 2018
CountryGermany
CityMunich
Period8/09/1814/09/18

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.

    Research areas

  • Adversarial synthesis, Deep compression, Face filtering, Person re-identification, Selective augmentation

Event

15th European Conference on Computer Vision, ECCV 2018

Duration8 Sep 201814 Sep 2018
CityMunich
CountryGermany

Event: Conference

Documents

Documents

  • Full-text PDF (accepted author manuscript)

    Rights statement: This is the accepted author manuscript (AAM). The final published version (version of record) is available online via Springer Link at https://doi.org/10.1007/978-3-030-11012-3_41 . Please refer to any applicable terms of use of the publisher.

    Accepted author manuscript, 1 MB, PDF document

    Embargo ends: 29/01/20

    Request copy

    Licence: Other

DOI

View research connections

Related faculties, schools or groups