Matches in SemOpenAlex for { <https://semopenalex.org/work/W3133672610> ?p ?o ?g. }
- W3133672610 abstract "Convolutional Neural Networks (CNNs) have proved very accurate in multiple computer vision image classification tasks that required visual inspection in the past (e.g., object recognition, face detection, etc.). Motivated by these astonishing results, researchers have also started using CNNs to cope with image forensic problems (e.g., camera model identification, tampering detection, etc.). However, in computer vision, image classification methods typically rely on visual cues easily detectable by human eyes. Conversely, forensic solutions rely on almost invisible traces that are often very subtle and lie in the fine details of the image under analysis. For this reason, training a CNN to solve a forensic task requires some special care, as common processing operations (e.g., resampling, compression, etc.) can strongly hinder forensic traces. In this work, we focus on the effect that JPEG has on CNN training considering different computer vision and forensic image classification problems. Specifically, we consider the issues that rise from JPEG compression and misalignment of the JPEG grid. We show that it is necessary to consider these effects when generating a training dataset in order to properly train a forensic detector not losing generalization capability, whereas it is almost possible to ignore these effects for computer vision tasks." @default.
- W3133672610 created "2021-03-15" @default.
- W3133672610 creator A5005378965 @default.
- W3133672610 creator A5043644234 @default.
- W3133672610 creator A5046125100 @default.
- W3133672610 creator A5051370303 @default.
- W3133672610 date "2020-12-06" @default.
- W3133672610 modified "2023-10-11" @default.
- W3133672610 title "Training CNNs in Presence of JPEG Compression: Multimedia Forensics vs Computer Vision" @default.
- W3133672610 cites W2050991548 @default.
- W3133672610 cites W2071018220 @default.
- W3133672610 cites W2108598243 @default.
- W3133672610 cites W2118779391 @default.
- W3133672610 cites W2130225759 @default.
- W3133672610 cites W2156619667 @default.
- W3133672610 cites W2194775991 @default.
- W3133672610 cites W2531409750 @default.
- W3133672610 cites W2545896937 @default.
- W3133672610 cites W2561004319 @default.
- W3133672610 cites W2572787276 @default.
- W3133672610 cites W2739273633 @default.
- W3133672610 cites W2743750471 @default.
- W3133672610 cites W2762911267 @default.
- W3133672610 cites W2787755382 @default.
- W3133672610 cites W2806156052 @default.
- W3133672610 cites W2888338418 @default.
- W3133672610 cites W2890052933 @default.
- W3133672610 cites W2890588722 @default.
- W3133672610 cites W2907295878 @default.
- W3133672610 cites W2911424785 @default.
- W3133672610 cites W2913323966 @default.
- W3133672610 cites W2913920880 @default.
- W3133672610 cites W2949446228 @default.
- W3133672610 cites W2962793481 @default.
- W3133672610 cites W2962866211 @default.
- W3133672610 cites W2980373331 @default.
- W3133672610 cites W3010699567 @default.
- W3133672610 cites W3011265505 @default.
- W3133672610 cites W3034530968 @default.
- W3133672610 cites W3034577585 @default.
- W3133672610 cites W3099319035 @default.
- W3133672610 cites W3102373040 @default.
- W3133672610 doi "https://doi.org/10.1109/wifs49906.2020.9360903" @default.
- W3133672610 hasPublicationYear "2020" @default.
- W3133672610 type Work @default.
- W3133672610 sameAs 3133672610 @default.
- W3133672610 citedByCount "16" @default.
- W3133672610 countsByYear W31336726102020 @default.
- W3133672610 countsByYear W31336726102021 @default.
- W3133672610 countsByYear W31336726102022 @default.
- W3133672610 countsByYear W31336726102023 @default.
- W3133672610 crossrefType "proceedings-article" @default.
- W3133672610 hasAuthorship W3133672610A5005378965 @default.
- W3133672610 hasAuthorship W3133672610A5043644234 @default.
- W3133672610 hasAuthorship W3133672610A5046125100 @default.
- W3133672610 hasAuthorship W3133672610A5051370303 @default.
- W3133672610 hasBestOaLocation W31336726102 @default.
- W3133672610 hasConcept C115961682 @default.
- W3133672610 hasConcept C13481523 @default.
- W3133672610 hasConcept C153180895 @default.
- W3133672610 hasConcept C154945302 @default.
- W3133672610 hasConcept C162324750 @default.
- W3133672610 hasConcept C187736073 @default.
- W3133672610 hasConcept C198751489 @default.
- W3133672610 hasConcept C2776151529 @default.
- W3133672610 hasConcept C2780451532 @default.
- W3133672610 hasConcept C31972630 @default.
- W3133672610 hasConcept C41008148 @default.
- W3133672610 hasConcept C69216139 @default.
- W3133672610 hasConcept C78548338 @default.
- W3133672610 hasConcept C81363708 @default.
- W3133672610 hasConcept C9417928 @default.
- W3133672610 hasConceptScore W3133672610C115961682 @default.
- W3133672610 hasConceptScore W3133672610C13481523 @default.
- W3133672610 hasConceptScore W3133672610C153180895 @default.
- W3133672610 hasConceptScore W3133672610C154945302 @default.
- W3133672610 hasConceptScore W3133672610C162324750 @default.
- W3133672610 hasConceptScore W3133672610C187736073 @default.
- W3133672610 hasConceptScore W3133672610C198751489 @default.
- W3133672610 hasConceptScore W3133672610C2776151529 @default.
- W3133672610 hasConceptScore W3133672610C2780451532 @default.
- W3133672610 hasConceptScore W3133672610C31972630 @default.
- W3133672610 hasConceptScore W3133672610C41008148 @default.
- W3133672610 hasConceptScore W3133672610C69216139 @default.
- W3133672610 hasConceptScore W3133672610C78548338 @default.
- W3133672610 hasConceptScore W3133672610C81363708 @default.
- W3133672610 hasConceptScore W3133672610C9417928 @default.
- W3133672610 hasFunder F4320338294 @default.
- W3133672610 hasLocation W31336726101 @default.
- W3133672610 hasLocation W31336726102 @default.
- W3133672610 hasLocation W31336726103 @default.
- W3133672610 hasOpenAccess W3133672610 @default.
- W3133672610 hasPrimaryLocation W31336726101 @default.
- W3133672610 hasRelatedWork W1968318766 @default.
- W3133672610 hasRelatedWork W2096442341 @default.
- W3133672610 hasRelatedWork W2159814397 @default.
- W3133672610 hasRelatedWork W2357838320 @default.
- W3133672610 hasRelatedWork W2441603424 @default.
- W3133672610 hasRelatedWork W2542654834 @default.
- W3133672610 hasRelatedWork W2971784143 @default.