NAM - Unsupervised cross-domain image mapping without cycles or GANs

פרסום מחקרי: תוצר מחקר מכנסהרצאהביקורת עמיתים

תקציר

Several methods were recently proposed for Unsupervised Domain Mapping, which is the task of translating images between domains without prior knowledge of correspondences. Current approaches suffer from an instability in training due to relying on GANs which are powerful but highly sensitive to hyper-parameters and suffer from mode collapse. In addition, most methods rely heavily on “cycle” relationships between the domains, which enforce a one-to-one mapping. In this work, we introduce an alternative method: NAM. NAM relies on a pre-trained generative model of the source domain, and aligns each target image with an image sampled from the source distribution while jointly optimizing the domain mapping function. Experiments are presented validating the effectiveness of our method.

שפה מקוריתאנגלית אמריקאית
סטטוס פרסוםפורסם - 2018
אירוע6th International Conference on Learning Representations, ICLR 2018 - Vancouver, קנדה
משך הזמן: 30 אפר׳ 20183 מאי 2018

כנס

כנס6th International Conference on Learning Representations, ICLR 2018
מדינה/אזורקנדה
עירVancouver
תקופה30/04/183/05/18

ASJC Scopus subject areas

  • ???subjectarea.asjc.3300.3304???
  • ???subjectarea.asjc.1700.1706???
  • ???subjectarea.asjc.3300.3310???
  • ???subjectarea.asjc.1200.1203???

טביעת אצבע

להלן מוצגים תחומי המחקר של הפרסום 'NAM - Unsupervised cross-domain image mapping without cycles or GANs'. יחד הם יוצרים טביעת אצבע ייחודית.

פורמט ציטוט ביבליוגרפי