FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963588253> ?p ?o ?g. }

• W2963588253 endingPage "1668" @default.
• W2963588253 startingPage "1655" @default.
• W2963588253 abstract "Image descriptors based on activations of Convolutional Neural Networks (CNNs) have become dominant in image retrieval due to their discriminative power, compactness of representation, and search efficiency. Training of CNNs, either from scratch or fine-tuning, requires a large amount of annotated data, where a high quality of annotation is often crucial. In this work, we propose to fine-tune CNNs for image retrieval on a large collection of unordered images in a fully automated manner. Reconstructed 3D models obtained by the state-of-the-art retrieval and structure-from-motion methods guide the selection of the training data. We show that both hard-positive and hard-negative examples, selected by exploiting the geometry and the camera positions available from the 3D models, enhance the performance of particular-object retrieval. CNN descriptor whitening discriminatively learned from the same training data outperforms commonly used PCA whitening. We propose a novel trainable Generalized-Mean (GeM) pooling layer that generalizes max and average pooling and show that it boosts retrieval performance. Applying the proposed method to the VGG network achieves state-of-the-art performance on the standard benchmarks: Oxford Buildings, Paris, and Holidays datasets." @default.
• W2963588253 created "2019-07-30" @default.
• W2963588253 creator A5019409717 @default.
• W2963588253 creator A5038939197 @default.
• W2963588253 creator A5046083819 @default.
• W2963588253 date "2019-07-01" @default.
• W2963588253 modified "2023-10-09" @default.
• W2963588253 title "Fine-Tuning CNN Image Retrieval with No Human Annotation" @default.
• W2963588253 cites W1524680991 @default.
• W2963588253 cites W1556531089 @default.
• W2963588253 cites W1565312575 @default.
• W2963588253 cites W1608697715 @default.
• W2963588253 cites W1908016767 @default.
• W2963588253 cites W1920702274 @default.
• W2963588253 cites W1926837203 @default.
• W2963588253 cites W1949049686 @default.
• W2963588253 cites W1963882359 @default.
• W2963588253 cites W1975517671 @default.
• W2963588253 cites W1979931042 @default.
• W2963588253 cites W1980911747 @default.
• W2963588253 cites W1984309565 @default.
• W2963588253 cites W2003340926 @default.
• W2963588253 cites W2007477183 @default.
• W2963588253 cites W2023991840 @default.
• W2963588253 cites W2024047694 @default.
• W2963588253 cites W204268067 @default.
• W2963588253 cites W2062118960 @default.
• W2963588253 cites W2071027807 @default.
• W2963588253 cites W2076434944 @default.
• W2963588253 cites W2086649520 @default.
• W2963588253 cites W2099443716 @default.
• W2963588253 cites W2099482512 @default.
• W2963588253 cites W2102605133 @default.
• W2963588253 cites W2103924867 @default.
• W2963588253 cites W2117539524 @default.
• W2963588253 cites W2138621090 @default.
• W2963588253 cites W2141362318 @default.
• W2963588253 cites W2143432975 @default.
• W2963588253 cites W2148809531 @default.
• W2963588253 cites W2149357475 @default.
• W2963588253 cites W2150307973 @default.
• W2963588253 cites W2157364932 @default.
• W2963588253 cites W2159386181 @default.
• W2963588253 cites W2161381512 @default.
• W2963588253 cites W2163446794 @default.
• W2963588253 cites W2194775991 @default.
• W2963588253 cites W2295537791 @default.
• W2963588253 cites W2299456531 @default.
• W2963588253 cites W2336803177 @default.
• W2963588253 cites W2340690086 @default.
• W2963588253 cites W2471962767 @default.
• W2963588253 cites W2544587078 @default.
• W2963588253 cites W2963125676 @default.
• W2963588253 cites W2963129433 @default.
• W2963588253 cites W2963775347 @default.
• W2963588253 cites W2964157791 @default.
• W2963588253 cites W3099206234 @default.
• W2963588253 cites W4236892273 @default.
• W2963588253 cites W54214434 @default.
• W2963588253 cites W56385144 @default.
• W2963588253 doi "https://doi.org/10.1109/tpami.2018.2846566" @default.
• W2963588253 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/29994246" @default.
• W2963588253 hasPublicationYear "2019" @default.
• W2963588253 type Work @default.
• W2963588253 sameAs 2963588253 @default.
• W2963588253 citedByCount "622" @default.
• W2963588253 countsByYear W29635882532018 @default.
• W2963588253 countsByYear W29635882532019 @default.
• W2963588253 countsByYear W29635882532020 @default.
• W2963588253 countsByYear W29635882532021 @default.
• W2963588253 countsByYear W29635882532022 @default.
• W2963588253 countsByYear W29635882532023 @default.
• W2963588253 crossrefType "journal-article" @default.
• W2963588253 hasAuthorship W2963588253A5019409717 @default.
• W2963588253 hasAuthorship W2963588253A5038939197 @default.
• W2963588253 hasAuthorship W2963588253A5046083819 @default.
• W2963588253 hasBestOaLocation W29635882532 @default.
• W2963588253 hasConcept C115961682 @default.
• W2963588253 hasConcept C153180895 @default.
• W2963588253 hasConcept C154945302 @default.
• W2963588253 hasConcept C1667742 @default.
• W2963588253 hasConcept C199579030 @default.
• W2963588253 hasConcept C2776151529 @default.
• W2963588253 hasConcept C2776321320 @default.
• W2963588253 hasConcept C31972630 @default.
• W2963588253 hasConcept C41008148 @default.
• W2963588253 hasConcept C70437156 @default.
• W2963588253 hasConcept C81363708 @default.
• W2963588253 hasConcept C97931131 @default.
• W2963588253 hasConceptScore W2963588253C115961682 @default.
• W2963588253 hasConceptScore W2963588253C153180895 @default.
• W2963588253 hasConceptScore W2963588253C154945302 @default.
• W2963588253 hasConceptScore W2963588253C1667742 @default.
• W2963588253 hasConceptScore W2963588253C199579030 @default.
• W2963588253 hasConceptScore W2963588253C2776151529 @default.
• W2963588253 hasConceptScore W2963588253C2776321320 @default.
• W2963588253 hasConceptScore W2963588253C31972630 @default.
• W2963588253 hasConceptScore W2963588253C41008148 @default.

• next