Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387579400> ?p ?o ?g. }
- W4387579400 abstract "This paper studies deep unsupervised hashing, which has attracted increasing attention in large-scale image retrieval. The majority of recent approaches usually reconstruct semantic similarity information, which then guides the hash code learning. However, they still fail to achieve satisfactory performance in reality due to two reasons. On the one hand, without accurate supervised information, these methods usually fail to produce independent and robust hash codes with semantics information well preserved, which may hinder effective image retrieval. On the other hand, due to discrete constraints, how to effectively optimize the hashing network in an end-to-end manner with small quantization errors remains to be a problem. To address these difficulties, we propose a novel unsupervised hashing method called HARR to learn discriminative and high-quality hash codes. To comprehensively explore semantic similarity structure, HARR adopts the Winner-Take-All (WTA) hash to model the similarity structure. Then similarity-preserving hash codes are learned under the reliable guidance of the reconstructed similarity structure. Additionally, we improve the quality of hash codes by a bit correlation reduction module, which forces the cross-correlation matrix between a batch of hash codes under different augmentations to approach the identity matrix. In this way, the generated hash bits are expected to be invariant to disturbances with minimal redundancy, which can be further interpreted as an instantiation of the Information Bottleneck principle. Finally, for effective hashing network training, we minimize the cosine distances between real-value network outputs and their binary codes for small quantization errors. Extensive experiments demonstrate the effectiveness of our proposed HARR." @default.
- W4387579400 created "2023-10-13" @default.
- W4387579400 creator A5019574552 @default.
- W4387579400 creator A5019972814 @default.
- W4387579400 creator A5024965898 @default.
- W4387579400 creator A5028336314 @default.
- W4387579400 creator A5030001981 @default.
- W4387579400 creator A5044318086 @default.
- W4387579400 creator A5073859117 @default.
- W4387579400 creator A5078385889 @default.
- W4387579400 date "2023-10-12" @default.
- W4387579400 modified "2023-10-14" @default.
- W4387579400 title "HARR: Learning Discriminative and High-quality Hash Codes for Image Retrieval" @default.
- W4387579400 cites W1861492603 @default.
- W4387579400 cites W1939575207 @default.
- W4387579400 cites W1970584764 @default.
- W4387579400 cites W1979605853 @default.
- W4387579400 cites W2007972815 @default.
- W4387579400 cites W2070222581 @default.
- W4387579400 cites W2073681337 @default.
- W4387579400 cites W2138621090 @default.
- W4387579400 cites W2155803963 @default.
- W4387579400 cites W2207555655 @default.
- W4387579400 cites W2282231212 @default.
- W4387579400 cites W2411707397 @default.
- W4387579400 cites W2461086877 @default.
- W4387579400 cites W2509619282 @default.
- W4387579400 cites W2766724094 @default.
- W4387579400 cites W2781821509 @default.
- W4387579400 cites W2880214242 @default.
- W4387579400 cites W2964158883 @default.
- W4387579400 cites W2970518483 @default.
- W4387579400 cites W2972421866 @default.
- W4387579400 cites W3008547073 @default.
- W4387579400 cites W3021893287 @default.
- W4387579400 cites W3034205626 @default.
- W4387579400 cites W3035524453 @default.
- W4387579400 cites W3035732386 @default.
- W4387579400 cites W3098365576 @default.
- W4387579400 cites W3099359379 @default.
- W4387579400 cites W3108274592 @default.
- W4387579400 cites W4210885000 @default.
- W4387579400 cites W4294733375 @default.
- W4387579400 cites W4304083183 @default.
- W4387579400 cites W4321021417 @default.
- W4387579400 cites W4322627163 @default.
- W4387579400 doi "https://doi.org/10.1145/3627162" @default.
- W4387579400 hasPublicationYear "2023" @default.
- W4387579400 type Work @default.
- W4387579400 citedByCount "0" @default.
- W4387579400 crossrefType "journal-article" @default.
- W4387579400 hasAuthorship W4387579400A5019574552 @default.
- W4387579400 hasAuthorship W4387579400A5019972814 @default.
- W4387579400 hasAuthorship W4387579400A5024965898 @default.
- W4387579400 hasAuthorship W4387579400A5028336314 @default.
- W4387579400 hasAuthorship W4387579400A5030001981 @default.
- W4387579400 hasAuthorship W4387579400A5044318086 @default.
- W4387579400 hasAuthorship W4387579400A5073859117 @default.
- W4387579400 hasAuthorship W4387579400A5078385889 @default.
- W4387579400 hasBestOaLocation W43875794001 @default.
- W4387579400 hasConcept C11413529 @default.
- W4387579400 hasConcept C116058348 @default.
- W4387579400 hasConcept C122907437 @default.
- W4387579400 hasConcept C124101348 @default.
- W4387579400 hasConcept C133667856 @default.
- W4387579400 hasConcept C138111711 @default.
- W4387579400 hasConcept C153180895 @default.
- W4387579400 hasConcept C154945302 @default.
- W4387579400 hasConcept C33923547 @default.
- W4387579400 hasConcept C38652104 @default.
- W4387579400 hasConcept C41008148 @default.
- W4387579400 hasConcept C48372109 @default.
- W4387579400 hasConcept C56856351 @default.
- W4387579400 hasConcept C63435697 @default.
- W4387579400 hasConcept C67388219 @default.
- W4387579400 hasConcept C80444323 @default.
- W4387579400 hasConcept C94375191 @default.
- W4387579400 hasConcept C99138194 @default.
- W4387579400 hasConceptScore W4387579400C11413529 @default.
- W4387579400 hasConceptScore W4387579400C116058348 @default.
- W4387579400 hasConceptScore W4387579400C122907437 @default.
- W4387579400 hasConceptScore W4387579400C124101348 @default.
- W4387579400 hasConceptScore W4387579400C133667856 @default.
- W4387579400 hasConceptScore W4387579400C138111711 @default.
- W4387579400 hasConceptScore W4387579400C153180895 @default.
- W4387579400 hasConceptScore W4387579400C154945302 @default.
- W4387579400 hasConceptScore W4387579400C33923547 @default.
- W4387579400 hasConceptScore W4387579400C38652104 @default.
- W4387579400 hasConceptScore W4387579400C41008148 @default.
- W4387579400 hasConceptScore W4387579400C48372109 @default.
- W4387579400 hasConceptScore W4387579400C56856351 @default.
- W4387579400 hasConceptScore W4387579400C63435697 @default.
- W4387579400 hasConceptScore W4387579400C67388219 @default.
- W4387579400 hasConceptScore W4387579400C80444323 @default.
- W4387579400 hasConceptScore W4387579400C94375191 @default.
- W4387579400 hasConceptScore W4387579400C99138194 @default.
- W4387579400 hasLocation W43875794001 @default.
- W4387579400 hasOpenAccess W4387579400 @default.
- W4387579400 hasPrimaryLocation W43875794001 @default.
- W4387579400 hasRelatedWork W2000284985 @default.