Matches in SemOpenAlex for { <https://semopenalex.org/work/W2965187816> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W2965187816 abstract "Neural networks need big annotated datasets for training. However, manual annotation can be too expensive or even unfeasible for certain tasks, like multi-person 2D pose estimation with severe occlusions. A remedy for this is synthetic data with perfect ground truth. Here we explore two variations of synthetic data for this challenging problem; a dataset with purely synthetic humans and a real dataset augmented with synthetic humans. We then study which approach better generalizes to real data, as well as the influence of virtual humans in the training loss. Using the augmented dataset, without considering synthetic humans in the loss, leads to the best results. We observe that not all synthetic samples are equally informative for training, while the informative samples are different for each training stage. To exploit this observation, we employ an adversarial student-teacher framework; the teacher improves the student by providing the hardest samples for its current state as a challenge. Experiments show that the student-teacher framework outperforms normal training on the purely synthetic dataset." @default.
- W2965187816 created "2019-08-13" @default.
- W2965187816 creator A5030083876 @default.
- W2965187816 creator A5056265728 @default.
- W2965187816 creator A5059854574 @default.
- W2965187816 creator A5070040247 @default.
- W2965187816 date "2019-08-02" @default.
- W2965187816 modified "2023-09-23" @default.
- W2965187816 title "Learning to Train with Synthetic Humans" @default.
- W2965187816 cites W1686810756 @default.
- W2965187816 cites W1943191679 @default.
- W2965187816 cites W197884394 @default.
- W2965187816 cites W2099333815 @default.
- W2965187816 cites W2122633688 @default.
- W2965187816 cites W2227547437 @default.
- W2965187816 cites W2559085405 @default.
- W2965187816 cites W2576289912 @default.
- W2965187816 cites W2768683308 @default.
- W2965187816 cites W2792747672 @default.
- W2965187816 cites W2963150697 @default.
- W2965187816 cites W2964121744 @default.
- W2965187816 cites W2964181100 @default.
- W2965187816 cites W2964238416 @default.
- W2965187816 cites W2997256263 @default.
- W2965187816 cites W76330425 @default.
- W2965187816 doi "https://doi.org/10.48550/arxiv.1908.00967" @default.
- W2965187816 hasPublicationYear "2019" @default.
- W2965187816 type Work @default.
- W2965187816 sameAs 2965187816 @default.
- W2965187816 citedByCount "0" @default.
- W2965187816 crossrefType "posted-content" @default.
- W2965187816 hasAuthorship W2965187816A5030083876 @default.
- W2965187816 hasAuthorship W2965187816A5056265728 @default.
- W2965187816 hasAuthorship W2965187816A5059854574 @default.
- W2965187816 hasAuthorship W2965187816A5070040247 @default.
- W2965187816 hasBestOaLocation W29651878161 @default.
- W2965187816 hasConcept C119857082 @default.
- W2965187816 hasConcept C146849305 @default.
- W2965187816 hasConcept C154945302 @default.
- W2965187816 hasConcept C160920958 @default.
- W2965187816 hasConcept C165696696 @default.
- W2965187816 hasConcept C2776321320 @default.
- W2965187816 hasConcept C37736160 @default.
- W2965187816 hasConcept C38652104 @default.
- W2965187816 hasConcept C41008148 @default.
- W2965187816 hasConcept C51632099 @default.
- W2965187816 hasConceptScore W2965187816C119857082 @default.
- W2965187816 hasConceptScore W2965187816C146849305 @default.
- W2965187816 hasConceptScore W2965187816C154945302 @default.
- W2965187816 hasConceptScore W2965187816C160920958 @default.
- W2965187816 hasConceptScore W2965187816C165696696 @default.
- W2965187816 hasConceptScore W2965187816C2776321320 @default.
- W2965187816 hasConceptScore W2965187816C37736160 @default.
- W2965187816 hasConceptScore W2965187816C38652104 @default.
- W2965187816 hasConceptScore W2965187816C41008148 @default.
- W2965187816 hasConceptScore W2965187816C51632099 @default.
- W2965187816 hasLocation W29651878161 @default.
- W2965187816 hasOpenAccess W2965187816 @default.
- W2965187816 hasPrimaryLocation W29651878161 @default.
- W2965187816 hasRelatedWork W2898291644 @default.
- W2965187816 hasRelatedWork W2963456518 @default.
- W2965187816 hasRelatedWork W3015448905 @default.
- W2965187816 hasRelatedWork W3033833039 @default.
- W2965187816 hasRelatedWork W3104224589 @default.
- W2965187816 hasRelatedWork W3124408655 @default.
- W2965187816 hasRelatedWork W4226226315 @default.
- W2965187816 hasRelatedWork W4283790566 @default.
- W2965187816 hasRelatedWork W4297785512 @default.
- W2965187816 hasRelatedWork W4297845793 @default.
- W2965187816 isParatext "false" @default.
- W2965187816 isRetracted "false" @default.
- W2965187816 magId "2965187816" @default.
- W2965187816 workType "article" @default.