Matches in SemOpenAlex for { <https://semopenalex.org/work/W3175255259> ?p ?o ?g. }
- W3175255259 abstract "Camera model identification is a standard task in digital image forensics. Learning-based approaches achieve state-of-the-art performance, but they are sensitive to so-called out-of-distribution (OOD) data due to a mismatch between the training and testing distribution. This may result in a significant reduction in classifier performance that is, unfortunately, not easy to anticipate for a forensic analyst.In this work, we investigate possibilities for adding reliability measures to the task of camera model identification. We leverage learning architectures that include an uncertainty measure with every prediction that can be reported back to an analyst. To this end, we investigate deep ensembles and Bayesian neural networks (BNNs). We compare both methods against a standard CNN with softmax statistics as uncertainty metric. We demonstrate in several experiments that both probabilistic approaches provide simultaneously state-of-the-art classification performance and reliable uncertainty estimates on OOD data. The uncertainty of deep ensembles is more accurate on OOD camera models, while BNN uncertainties are more accurate on OOD post-processing." @default.
- W3175255259 created "2021-07-05" @default.
- W3175255259 creator A5008546345 @default.
- W3175255259 creator A5039525758 @default.
- W3175255259 creator A5049250339 @default.
- W3175255259 creator A5073442830 @default.
- W3175255259 date "2021-05-06" @default.
- W3175255259 modified "2023-09-24" @default.
- W3175255259 title "Reliable Camera Model Identification Through Uncertainty Estimation" @default.
- W3175255259 cites W2047229728 @default.
- W3175255259 cites W2084252081 @default.
- W3175255259 cites W2095669091 @default.
- W3175255259 cites W2108677974 @default.
- W3175255259 cites W2124929806 @default.
- W3175255259 cites W2151939419 @default.
- W3175255259 cites W2164411961 @default.
- W3175255259 cites W2531327146 @default.
- W3175255259 cites W2572787276 @default.
- W3175255259 cites W2600383743 @default.
- W3175255259 cites W2798117183 @default.
- W3175255259 cites W2802701183 @default.
- W3175255259 cites W2888338418 @default.
- W3175255259 cites W2890588722 @default.
- W3175255259 cites W2900420778 @default.
- W3175255259 cites W2902304528 @default.
- W3175255259 cites W2938564271 @default.
- W3175255259 cites W2963238274 @default.
- W3175255259 cites W2963351448 @default.
- W3175255259 cites W2963693742 @default.
- W3175255259 cites W2963939004 @default.
- W3175255259 cites W2964059111 @default.
- W3175255259 cites W2964212410 @default.
- W3175255259 cites W2965563166 @default.
- W3175255259 cites W2970859221 @default.
- W3175255259 cites W2972328267 @default.
- W3175255259 cites W2982058372 @default.
- W3175255259 cites W3089377254 @default.
- W3175255259 cites W3133672610 @default.
- W3175255259 cites W3135322301 @default.
- W3175255259 cites W3135434620 @default.
- W3175255259 doi "https://doi.org/10.1109/iwbf50991.2021.9465097" @default.
- W3175255259 hasPublicationYear "2021" @default.
- W3175255259 type Work @default.
- W3175255259 sameAs 3175255259 @default.
- W3175255259 citedByCount "0" @default.
- W3175255259 crossrefType "proceedings-article" @default.
- W3175255259 hasAuthorship W3175255259A5008546345 @default.
- W3175255259 hasAuthorship W3175255259A5039525758 @default.
- W3175255259 hasAuthorship W3175255259A5049250339 @default.
- W3175255259 hasAuthorship W3175255259A5073442830 @default.
- W3175255259 hasConcept C108583219 @default.
- W3175255259 hasConcept C116834253 @default.
- W3175255259 hasConcept C119857082 @default.
- W3175255259 hasConcept C124101348 @default.
- W3175255259 hasConcept C127413603 @default.
- W3175255259 hasConcept C153083717 @default.
- W3175255259 hasConcept C153180895 @default.
- W3175255259 hasConcept C154945302 @default.
- W3175255259 hasConcept C176217482 @default.
- W3175255259 hasConcept C188441871 @default.
- W3175255259 hasConcept C21547014 @default.
- W3175255259 hasConcept C41008148 @default.
- W3175255259 hasConcept C49937458 @default.
- W3175255259 hasConcept C50644808 @default.
- W3175255259 hasConcept C59822182 @default.
- W3175255259 hasConcept C81363708 @default.
- W3175255259 hasConcept C86803240 @default.
- W3175255259 hasConcept C95623464 @default.
- W3175255259 hasConceptScore W3175255259C108583219 @default.
- W3175255259 hasConceptScore W3175255259C116834253 @default.
- W3175255259 hasConceptScore W3175255259C119857082 @default.
- W3175255259 hasConceptScore W3175255259C124101348 @default.
- W3175255259 hasConceptScore W3175255259C127413603 @default.
- W3175255259 hasConceptScore W3175255259C153083717 @default.
- W3175255259 hasConceptScore W3175255259C153180895 @default.
- W3175255259 hasConceptScore W3175255259C154945302 @default.
- W3175255259 hasConceptScore W3175255259C176217482 @default.
- W3175255259 hasConceptScore W3175255259C188441871 @default.
- W3175255259 hasConceptScore W3175255259C21547014 @default.
- W3175255259 hasConceptScore W3175255259C41008148 @default.
- W3175255259 hasConceptScore W3175255259C49937458 @default.
- W3175255259 hasConceptScore W3175255259C50644808 @default.
- W3175255259 hasConceptScore W3175255259C59822182 @default.
- W3175255259 hasConceptScore W3175255259C81363708 @default.
- W3175255259 hasConceptScore W3175255259C86803240 @default.
- W3175255259 hasConceptScore W3175255259C95623464 @default.
- W3175255259 hasFunder F4320321408 @default.
- W3175255259 hasLocation W31752552591 @default.
- W3175255259 hasOpenAccess W3175255259 @default.
- W3175255259 hasPrimaryLocation W31752552591 @default.
- W3175255259 hasRelatedWork W11389402 @default.
- W3175255259 hasRelatedWork W14516383 @default.
- W3175255259 hasRelatedWork W2701911 @default.
- W3175255259 hasRelatedWork W3370582 @default.
- W3175255259 hasRelatedWork W482721 @default.
- W3175255259 hasRelatedWork W5641948 @default.
- W3175255259 hasRelatedWork W6680660 @default.
- W3175255259 hasRelatedWork W7965686 @default.
- W3175255259 hasRelatedWork W8198582 @default.
- W3175255259 hasRelatedWork W9190101 @default.