Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385800690> ?p ?o ?g. }
- W4385800690 abstract "Image anonymization is widely adapted in practice to comply with privacy regulations in many regions. However, anonymization often degrades the quality of the data, reducing its utility for computer vision development. In this paper, we investigate the impact of image anonymization for training computer vision models on key computer vision tasks (detection, instance segmentation, and pose estimation). Specifically, we benchmark the recognition drop on common detection datasets, where we evaluate both traditional and realistic anonymization for faces and full bodies. Our comprehensive experiments reflect that traditional image anonymization substantially impacts final model performance, particularly when anonymizing the full body. Furthermore, we find that realistic anonymization can mitigate this decrease in performance, where our experiments reflect a minimal performance drop for face anonymization. Our study demonstrates that realistic anonymization can enable privacy-preserving computer vision development with minimal performance degradation across a range of important computer vision benchmarks." @default.
- W4385800690 created "2023-08-15" @default.
- W4385800690 creator A5002687932 @default.
- W4385800690 creator A5038367345 @default.
- W4385800690 date "2023-06-01" @default.
- W4385800690 modified "2023-10-16" @default.
- W4385800690 title "Does Image Anonymization Impact Computer Vision Training?" @default.
- W4385800690 cites W1488228643 @default.
- W4385800690 cites W1523682477 @default.
- W4385800690 cites W1892323599 @default.
- W4385800690 cites W1969513994 @default.
- W4385800690 cites W1972138918 @default.
- W4385800690 cites W2055860749 @default.
- W4385800690 cites W2103958416 @default.
- W4385800690 cites W2108598243 @default.
- W4385800690 cites W2116907524 @default.
- W4385800690 cites W2128354257 @default.
- W4385800690 cites W2156001867 @default.
- W4385800690 cites W2158551678 @default.
- W4385800690 cites W2194775991 @default.
- W4385800690 cites W2285252196 @default.
- W4385800690 cites W2340897893 @default.
- W4385800690 cites W2480709345 @default.
- W4385800690 cites W2528608753 @default.
- W4385800690 cites W2605289356 @default.
- W4385800690 cites W2741333573 @default.
- W4385800690 cites W2747472649 @default.
- W4385800690 cites W2769823506 @default.
- W4385800690 cites W2774833025 @default.
- W4385800690 cites W2795931522 @default.
- W4385800690 cites W2797132229 @default.
- W4385800690 cites W2914620083 @default.
- W4385800690 cites W2949924060 @default.
- W4385800690 cites W2962712356 @default.
- W4385800690 cites W2962766044 @default.
- W4385800690 cites W2962770929 @default.
- W4385800690 cites W2963150697 @default.
- W4385800690 cites W2963542245 @default.
- W4385800690 cites W2963566548 @default.
- W4385800690 cites W2963876278 @default.
- W4385800690 cites W2982340830 @default.
- W4385800690 cites W3035211587 @default.
- W4385800690 cites W3035564946 @default.
- W4385800690 cites W3174879512 @default.
- W4385800690 cites W3207351435 @default.
- W4385800690 cites W4210327842 @default.
- W4385800690 cites W4285981707 @default.
- W4385800690 cites W4286285644 @default.
- W4385800690 cites W4319300095 @default.
- W4385800690 cites W4319300278 @default.
- W4385800690 cites W4319300743 @default.
- W4385800690 doi "https://doi.org/10.1109/cvprw59228.2023.00019" @default.
- W4385800690 hasPublicationYear "2023" @default.
- W4385800690 type Work @default.
- W4385800690 citedByCount "0" @default.
- W4385800690 crossrefType "proceedings-article" @default.
- W4385800690 hasAuthorship W4385800690A5002687932 @default.
- W4385800690 hasAuthorship W4385800690A5038367345 @default.
- W4385800690 hasBestOaLocation W43858006902 @default.
- W4385800690 hasConcept C115961682 @default.
- W4385800690 hasConcept C124101348 @default.
- W4385800690 hasConcept C13280743 @default.
- W4385800690 hasConcept C144024400 @default.
- W4385800690 hasConcept C153180895 @default.
- W4385800690 hasConcept C154945302 @default.
- W4385800690 hasConcept C185798385 @default.
- W4385800690 hasConcept C205649164 @default.
- W4385800690 hasConcept C26517878 @default.
- W4385800690 hasConcept C2779304628 @default.
- W4385800690 hasConcept C31510193 @default.
- W4385800690 hasConcept C31972630 @default.
- W4385800690 hasConcept C36289849 @default.
- W4385800690 hasConcept C38652104 @default.
- W4385800690 hasConcept C41008148 @default.
- W4385800690 hasConcept C89600930 @default.
- W4385800690 hasConceptScore W4385800690C115961682 @default.
- W4385800690 hasConceptScore W4385800690C124101348 @default.
- W4385800690 hasConceptScore W4385800690C13280743 @default.
- W4385800690 hasConceptScore W4385800690C144024400 @default.
- W4385800690 hasConceptScore W4385800690C153180895 @default.
- W4385800690 hasConceptScore W4385800690C154945302 @default.
- W4385800690 hasConceptScore W4385800690C185798385 @default.
- W4385800690 hasConceptScore W4385800690C205649164 @default.
- W4385800690 hasConceptScore W4385800690C26517878 @default.
- W4385800690 hasConceptScore W4385800690C2779304628 @default.
- W4385800690 hasConceptScore W4385800690C31510193 @default.
- W4385800690 hasConceptScore W4385800690C31972630 @default.
- W4385800690 hasConceptScore W4385800690C36289849 @default.
- W4385800690 hasConceptScore W4385800690C38652104 @default.
- W4385800690 hasConceptScore W4385800690C41008148 @default.
- W4385800690 hasConceptScore W4385800690C89600930 @default.
- W4385800690 hasLocation W43858006901 @default.
- W4385800690 hasLocation W43858006902 @default.
- W4385800690 hasOpenAccess W4385800690 @default.
- W4385800690 hasPrimaryLocation W43858006901 @default.
- W4385800690 hasRelatedWork W1669643531 @default.
- W4385800690 hasRelatedWork W1775397219 @default.
- W4385800690 hasRelatedWork W2005437358 @default.
- W4385800690 hasRelatedWork W2008656436 @default.
- W4385800690 hasRelatedWork W2023558673 @default.