?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Deep+Compact+Person+Re-Identification+with+Distractor+Synthesis+via+Guided+DC-GANs&rft.creator=Ponce-Lopez%2C+V&rft.creator=Burghardt%2C+T&rft.creator=Sun%2C+Y&rft.creator=Hannuna%2C+S&rft.creator=Damen%2C+D&rft.creator=Mirmehdi%2C+M&rft.description=We+present+a+dual-stream+CNN+that+learns+both+appearance+and+facial+features+in+tandem+from+still+images+and%2C+after+feature+fusion%2C+infers+person+identities.+We+then+describe+an+alternative+architecture+of+a+single%2C+lightweight+ID-CondenseNet+where+a+face+detector-guided+DC-GAN+is+used+to+generate+distractor+person+images+for+enhanced+training.+For+evaluation%2C+we+test+both+architectures+on+FLIMA%2C+a+new+extension+of+an+existing+person+re-identification+dataset+with+added+frame-by-frame+annotations+of+face+presence.+Although+the+dual-stream+CNN+can+outperform+the+CondenseNet+approach+on+FLIMA%2C+we+show+that+the+latter+surpasses+all+state-of-the-art+architectures+in+top-1+ranking+performance+when+applied+to+the+largest+existing+person+re-identification+dataset%2C+MSMT17.+We+conclude+that+whilst+re-identification+performance+is+highly+sensitive+to+the+structure+of+datasets%2C+distractor+augmentation+and+network+compression+have+a+role+to+play+for+enhancing+performance+characteristics+for+larger+scale+applications.&rft.subject=Person+Re-ID%2C+GANs%2C+Distractor+synthesis%2C+Deep+face+analysis&rft.publisher=Springer&rft.contributor=Ricci%2C+E&rft.contributor=Bulo%2C+SR&rft.contributor=Snoek%2C+C&rft.contributor=Lanz%2C+O&rft.contributor=Messelodi%2C+S&rft.contributor=Sebe%2C+N&rft.date=2019-09-02&rft.type=Proceedings+paper&rft.publisher=20th+International+Conference+on+Image+Analysis+and+Processing+(ICIAP)&rft.language=eng&rft.source=+++++In%3A+Ricci%2C+E+and+Bulo%2C+SR+and+Snoek%2C+C+and+Lanz%2C+O+and+Messelodi%2C+S+and+Sebe%2C+N%2C+(eds.)+Image+Analysis+and+Processing+%E2%80%93+ICIAP+2019.+ICIAP+2019.++(pp.+pp.+488-498).++Springer%3A+Cham%2C+Switzerland.+(2019)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10115347%2F1%2FPonce%2520Lopez_ICIAP19_Guided-DC-GANs_CameraReady.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10115347%2F&rft.rights=open