Matches in SemOpenAlex for { <https://semopenalex.org/work/W1869500417> ?p ?o ?g. }
- W1869500417 abstract "Deep learning has revolutionalized image-level tasks such as classification, but patch-level tasks, such as correspondence, still rely on hand-crafted features, e.g. SIFT. In this paper we use Convolutional Neural Networks (CNNs) to learn discriminant patch representations and in particular train a Siamese network with pairs of (non-)corresponding patches. We deal with the large number of potential pairs with the combination of a stochastic sampling of the training set and an aggressive mining strategy biased towards patches that are hard to classify. By using the L2 distance during both training and testing we develop 128-D descriptors whose euclidean distances reflect patch similarity, and which can be used as a drop-in replacement for any task involving SIFT. We demonstrate consistent performance gains over the state of the art, and generalize well against scaling and rotation, perspective transformation, non-rigid deformation, and illumination changes. Our descriptors are efficient to compute and amenable to modern GPUs, and are publicly available." @default.
- W1869500417 created "2016-06-24" @default.
- W1869500417 creator A5005762795 @default.
- W1869500417 creator A5038674741 @default.
- W1869500417 creator A5039307249 @default.
- W1869500417 creator A5049610895 @default.
- W1869500417 creator A5061430475 @default.
- W1869500417 creator A5077035803 @default.
- W1869500417 date "2015-12-01" @default.
- W1869500417 modified "2023-10-07" @default.
- W1869500417 title "Discriminative Learning of Deep Convolutional Feature Point Descriptors" @default.
- W1869500417 cites W1677409904 @default.
- W1869500417 cites W1955055330 @default.
- W1869500417 cites W1976526581 @default.
- W1869500417 cites W2007178811 @default.
- W1869500417 cites W2033743519 @default.
- W1869500417 cites W2089888558 @default.
- W1869500417 cites W2104853049 @default.
- W1869500417 cites W2115362620 @default.
- W1869500417 cites W2117228865 @default.
- W1869500417 cites W2126338221 @default.
- W1869500417 cites W2127589108 @default.
- W1869500417 cites W2134514757 @default.
- W1869500417 cites W2144041313 @default.
- W1869500417 cites W2145038566 @default.
- W1869500417 cites W2151103935 @default.
- W1869500417 cites W2158893304 @default.
- W1869500417 cites W2168356304 @default.
- W1869500417 cites W2177274842 @default.
- W1869500417 cites W2268446623 @default.
- W1869500417 cites W2546302380 @default.
- W1869500417 doi "https://doi.org/10.1109/iccv.2015.22" @default.
- W1869500417 hasPublicationYear "2015" @default.
- W1869500417 type Work @default.
- W1869500417 sameAs 1869500417 @default.
- W1869500417 citedByCount "624" @default.
- W1869500417 countsByYear W18695004172014 @default.
- W1869500417 countsByYear W18695004172015 @default.
- W1869500417 countsByYear W18695004172016 @default.
- W1869500417 countsByYear W18695004172017 @default.
- W1869500417 countsByYear W18695004172018 @default.
- W1869500417 countsByYear W18695004172019 @default.
- W1869500417 countsByYear W18695004172020 @default.
- W1869500417 countsByYear W18695004172021 @default.
- W1869500417 countsByYear W18695004172022 @default.
- W1869500417 countsByYear W18695004172023 @default.
- W1869500417 crossrefType "proceedings-article" @default.
- W1869500417 hasAuthorship W1869500417A5005762795 @default.
- W1869500417 hasAuthorship W1869500417A5038674741 @default.
- W1869500417 hasAuthorship W1869500417A5039307249 @default.
- W1869500417 hasAuthorship W1869500417A5049610895 @default.
- W1869500417 hasAuthorship W1869500417A5061430475 @default.
- W1869500417 hasAuthorship W1869500417A5077035803 @default.
- W1869500417 hasBestOaLocation W18695004173 @default.
- W1869500417 hasConcept C103278499 @default.
- W1869500417 hasConcept C108583219 @default.
- W1869500417 hasConcept C115961682 @default.
- W1869500417 hasConcept C120174047 @default.
- W1869500417 hasConcept C138885662 @default.
- W1869500417 hasConcept C153180895 @default.
- W1869500417 hasConcept C154945302 @default.
- W1869500417 hasConcept C177264268 @default.
- W1869500417 hasConcept C199360897 @default.
- W1869500417 hasConcept C2524010 @default.
- W1869500417 hasConcept C2776401178 @default.
- W1869500417 hasConcept C2779597229 @default.
- W1869500417 hasConcept C33923547 @default.
- W1869500417 hasConcept C41008148 @default.
- W1869500417 hasConcept C41895202 @default.
- W1869500417 hasConcept C52622490 @default.
- W1869500417 hasConcept C61265191 @default.
- W1869500417 hasConcept C81363708 @default.
- W1869500417 hasConcept C97931131 @default.
- W1869500417 hasConcept C99844830 @default.
- W1869500417 hasConceptScore W1869500417C103278499 @default.
- W1869500417 hasConceptScore W1869500417C108583219 @default.
- W1869500417 hasConceptScore W1869500417C115961682 @default.
- W1869500417 hasConceptScore W1869500417C120174047 @default.
- W1869500417 hasConceptScore W1869500417C138885662 @default.
- W1869500417 hasConceptScore W1869500417C153180895 @default.
- W1869500417 hasConceptScore W1869500417C154945302 @default.
- W1869500417 hasConceptScore W1869500417C177264268 @default.
- W1869500417 hasConceptScore W1869500417C199360897 @default.
- W1869500417 hasConceptScore W1869500417C2524010 @default.
- W1869500417 hasConceptScore W1869500417C2776401178 @default.
- W1869500417 hasConceptScore W1869500417C2779597229 @default.
- W1869500417 hasConceptScore W1869500417C33923547 @default.
- W1869500417 hasConceptScore W1869500417C41008148 @default.
- W1869500417 hasConceptScore W1869500417C41895202 @default.
- W1869500417 hasConceptScore W1869500417C52622490 @default.
- W1869500417 hasConceptScore W1869500417C61265191 @default.
- W1869500417 hasConceptScore W1869500417C81363708 @default.
- W1869500417 hasConceptScore W1869500417C97931131 @default.
- W1869500417 hasConceptScore W1869500417C99844830 @default.
- W1869500417 hasLocation W18695004171 @default.
- W1869500417 hasLocation W18695004172 @default.
- W1869500417 hasLocation W18695004173 @default.
- W1869500417 hasLocation W18695004174 @default.
- W1869500417 hasLocation W18695004175 @default.
- W1869500417 hasLocation W18695004176 @default.