Re-ID done right: towards good practices for person re-identification

Autor: Almazan, Jon, Gajic, Bojana, Murray, Naila, Larlus, Diane

Training a deep architecture using a ranking loss has become standard for the person re-identification task. Increasingly, these deep architectures include additional components that leverage part detections, attribute predictions, pose estimators an

Externí odkaz: http://arxiv.org/abs/1801.05339

Training strategies for efficient deep image retrieval

Autor: Gajić, Bojana

Publikováno v: Dipòsit Digital de Documents de la UAB
Universitat Autònoma de Barcelona
TDX (Tesis Doctorals en Xarxa)
TDR. Tesis Doctorales en Red
instname

En aquesta tesi ens centrem en la recuperació i re-identificació d’imatges. L’entrenament de xarxes neuronals profundes usant funcions de pèrdua basades en rànquing ha esdevingut un estàndard de facto per a les tasques de recuperació i re-i

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=dedup_wf_001::c71a6e1cb843ae0427df3690ccf47e8e
https://ddd.uab.cat/record/257619

Feature Extraction for Automatic Speech Recognition in Noisy Acoustic Environments

Autor: Gajic, Bojana

This thesis presents a study of alternative speech feature extraction methods aimed at increasing robustness of automatic speech recognition (ASR) against additive background noise. Spectral peak positions of speech signals remain practically unchang

Externí odkaz: http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-441