Matches in SemOpenAlex for { <https://semopenalex.org/work/W2791396492> ?p ?o ?g. }
- W2791396492 abstract "Deep supervised hashing has emerged as an influential solution to large-scale semantic image retrieval problems in computer vision. In the light of recent progress, convolutional neural network based hashing methods typically seek pair-wise or triplet labels to conduct the similarity preserving learning. However, complex semantic concepts of visual contents are hard to capture by similar/dissimilar labels, which limits the retrieval performance. Generally, pair-wise or triplet losses not only suffer from expensive training costs but also lack in extracting sufficient semantic information. In this regard, we propose a novel deep supervised hashing model to learn more compact class-level similarity preserving binary codes. Our deep learning based model is motivated by deep metric learning that directly takes semantic labels as supervised information in training and generates corresponding discriminant hashing code. Specifically, a novel cubic constraint loss function based on Gaussian distribution is proposed, which preserves semantic variations while penalizes the overlap part of different classes in the embedding space. To address the discrete optimization problem introduced by binary codes, a two-step optimization strategy is proposed to provide efficient training and avoid the problem of gradient vanishing. Extensive experiments on four large-scale benchmark databases show that our model can achieve the state-of-the-art retrieval performance. Moreover, when training samples are limited, our method surpasses other supervised deep hashing methods with non-negligible margins." @default.
- W2791396492 created "2018-03-29" @default.
- W2791396492 creator A5005150832 @default.
- W2791396492 creator A5019433605 @default.
- W2791396492 creator A5044236581 @default.
- W2791396492 date "2018-03-12" @default.
- W2791396492 modified "2023-10-16" @default.
- W2791396492 title "Deep Class-Wise Hashing: Semantics-Preserving Hashing via Class-wise Loss" @default.
- W2791396492 cites W1566135517 @default.
- W2791396492 cites W1861492603 @default.
- W2791396492 cites W1907458443 @default.
- W2791396492 cites W1939575207 @default.
- W2791396492 cites W1974647172 @default.
- W2791396492 cites W2023200581 @default.
- W2791396492 cites W2069870183 @default.
- W2791396492 cites W2117539524 @default.
- W2791396492 cites W2124386111 @default.
- W2791396492 cites W2144892774 @default.
- W2791396492 cites W2145072179 @default.
- W2791396492 cites W2161969291 @default.
- W2791396492 cites W2186615578 @default.
- W2791396492 cites W2187089797 @default.
- W2791396492 cites W2188403848 @default.
- W2791396492 cites W2270409809 @default.
- W2791396492 cites W2293824885 @default.
- W2791396492 cites W2411707397 @default.
- W2791396492 cites W2510911086 @default.
- W2791396492 cites W2520774990 @default.
- W2791396492 cites W2531549126 @default.
- W2791396492 cites W2555897561 @default.
- W2791396492 cites W2565993688 @default.
- W2791396492 cites W2586811659 @default.
- W2791396492 cites W2586937979 @default.
- W2791396492 cites W2594088761 @default.
- W2791396492 cites W2604880013 @default.
- W2791396492 cites W2606348288 @default.
- W2791396492 cites W2621176975 @default.
- W2791396492 cites W2740030094 @default.
- W2791396492 cites W2741824037 @default.
- W2791396492 cites W2949117887 @default.
- W2791396492 cites W2951581059 @default.
- W2791396492 cites W2962798895 @default.
- W2791396492 cites W2963026686 @default.
- W2791396492 cites W2963398644 @default.
- W2791396492 cites W2964076257 @default.
- W2791396492 cites W3118608800 @default.
- W2791396492 doi "https://doi.org/10.48550/arxiv.1803.04137" @default.
- W2791396492 hasPublicationYear "2018" @default.
- W2791396492 type Work @default.
- W2791396492 sameAs 2791396492 @default.
- W2791396492 citedByCount "0" @default.
- W2791396492 crossrefType "posted-content" @default.
- W2791396492 hasAuthorship W2791396492A5005150832 @default.
- W2791396492 hasAuthorship W2791396492A5019433605 @default.
- W2791396492 hasAuthorship W2791396492A5044236581 @default.
- W2791396492 hasBestOaLocation W27913964921 @default.
- W2791396492 hasConcept C108583219 @default.
- W2791396492 hasConcept C115961682 @default.
- W2791396492 hasConcept C116058348 @default.
- W2791396492 hasConcept C119857082 @default.
- W2791396492 hasConcept C130318100 @default.
- W2791396492 hasConcept C138111711 @default.
- W2791396492 hasConcept C153180895 @default.
- W2791396492 hasConcept C154945302 @default.
- W2791396492 hasConcept C1667742 @default.
- W2791396492 hasConcept C184337299 @default.
- W2791396492 hasConcept C199360897 @default.
- W2791396492 hasConcept C33923547 @default.
- W2791396492 hasConcept C38652104 @default.
- W2791396492 hasConcept C41008148 @default.
- W2791396492 hasConcept C48372109 @default.
- W2791396492 hasConcept C63435697 @default.
- W2791396492 hasConcept C67388219 @default.
- W2791396492 hasConcept C74270461 @default.
- W2791396492 hasConcept C81363708 @default.
- W2791396492 hasConcept C94375191 @default.
- W2791396492 hasConcept C99138194 @default.
- W2791396492 hasConceptScore W2791396492C108583219 @default.
- W2791396492 hasConceptScore W2791396492C115961682 @default.
- W2791396492 hasConceptScore W2791396492C116058348 @default.
- W2791396492 hasConceptScore W2791396492C119857082 @default.
- W2791396492 hasConceptScore W2791396492C130318100 @default.
- W2791396492 hasConceptScore W2791396492C138111711 @default.
- W2791396492 hasConceptScore W2791396492C153180895 @default.
- W2791396492 hasConceptScore W2791396492C154945302 @default.
- W2791396492 hasConceptScore W2791396492C1667742 @default.
- W2791396492 hasConceptScore W2791396492C184337299 @default.
- W2791396492 hasConceptScore W2791396492C199360897 @default.
- W2791396492 hasConceptScore W2791396492C33923547 @default.
- W2791396492 hasConceptScore W2791396492C38652104 @default.
- W2791396492 hasConceptScore W2791396492C41008148 @default.
- W2791396492 hasConceptScore W2791396492C48372109 @default.
- W2791396492 hasConceptScore W2791396492C63435697 @default.
- W2791396492 hasConceptScore W2791396492C67388219 @default.
- W2791396492 hasConceptScore W2791396492C74270461 @default.
- W2791396492 hasConceptScore W2791396492C81363708 @default.
- W2791396492 hasConceptScore W2791396492C94375191 @default.
- W2791396492 hasConceptScore W2791396492C99138194 @default.
- W2791396492 hasLocation W27913964921 @default.
- W2791396492 hasOpenAccess W2791396492 @default.