Matches in SemOpenAlex for { <https://semopenalex.org/work/W3188008942> ?p ?o ?g. }
- W3188008942 abstract "How do the neural networks distinguish two images? It is of critical importance to understand the matching mechanism of deep models for developing reliable intelligent systems for many risky visual applications such as surveillance and access control. However, most existing deep metric learning methods match the images by comparing feature vectors, which ignores the spatial structure of images and thus lacks interpretability. In this paper, we present a deep interpretable metric learning (DIML) method for more transparent embedding learning. Unlike conventional metric learning methods based on feature vector comparison, we propose a structural matching strategy that explicitly aligns the spatial embeddings by computing an optimal matching flow between feature maps of the two images. Our method enables deep models to learn metrics in a more human-friendly way, where the similarity of two images can be decomposed to several part-wise similarities and their contributions to the overall similarity. Our method is model-agnostic, which can be applied to off-the-shelf backbone networks and metric learning methods. We evaluate our method on three major benchmarks of deep metric learning including CUB200-2011, Cars196, and Stanford Online Products, and achieve substantial improvements over popular metric learning methods with better interpretability. Code is available at https://github.com/wl-zhao/DIML." @default.
- W3188008942 created "2021-08-16" @default.
- W3188008942 creator A5010357390 @default.
- W3188008942 creator A5060902476 @default.
- W3188008942 creator A5076637933 @default.
- W3188008942 creator A5082694832 @default.
- W3188008942 creator A5090079801 @default.
- W3188008942 date "2021-10-01" @default.
- W3188008942 modified "2023-10-16" @default.
- W3188008942 title "Towards Interpretable Deep Metric Learning with Structural Matching" @default.
- W3188008942 cites W2097117768 @default.
- W3188008942 cites W2138621090 @default.
- W3188008942 cites W2145287260 @default.
- W3188008942 cites W2194775991 @default.
- W3188008942 cites W2295107390 @default.
- W3188008942 cites W2605102252 @default.
- W3188008942 cites W2606377603 @default.
- W3188008942 cites W2764024122 @default.
- W3188008942 cites W2776507577 @default.
- W3188008942 cites W2885563408 @default.
- W3188008942 cites W2902512195 @default.
- W3188008942 cites W2910603373 @default.
- W3188008942 cites W2919115771 @default.
- W3188008942 cites W2948429466 @default.
- W3188008942 cites W2962858109 @default.
- W3188008942 cites W2962869940 @default.
- W3188008942 cites W2963026686 @default.
- W3188008942 cites W2963150697 @default.
- W3188008942 cites W2963350250 @default.
- W3188008942 cites W2963988212 @default.
- W3188008942 cites W2964271799 @default.
- W3188008942 cites W2969985801 @default.
- W3188008942 cites W2972578900 @default.
- W3188008942 cites W2979997102 @default.
- W3188008942 cites W2991581349 @default.
- W3188008942 cites W3034202663 @default.
- W3188008942 cites W3035102141 @default.
- W3188008942 cites W3035295689 @default.
- W3188008942 cites W3099206234 @default.
- W3188008942 cites W3110182042 @default.
- W3188008942 cites W3176196997 @default.
- W3188008942 doi "https://doi.org/10.1109/iccv48922.2021.00974" @default.
- W3188008942 hasPublicationYear "2021" @default.
- W3188008942 type Work @default.
- W3188008942 sameAs 3188008942 @default.
- W3188008942 citedByCount "11" @default.
- W3188008942 countsByYear W31880089422022 @default.
- W3188008942 countsByYear W31880089422023 @default.
- W3188008942 crossrefType "proceedings-article" @default.
- W3188008942 hasAuthorship W3188008942A5010357390 @default.
- W3188008942 hasAuthorship W3188008942A5060902476 @default.
- W3188008942 hasAuthorship W3188008942A5076637933 @default.
- W3188008942 hasAuthorship W3188008942A5082694832 @default.
- W3188008942 hasAuthorship W3188008942A5090079801 @default.
- W3188008942 hasBestOaLocation W31880089422 @default.
- W3188008942 hasConcept C103278499 @default.
- W3188008942 hasConcept C105795698 @default.
- W3188008942 hasConcept C108583219 @default.
- W3188008942 hasConcept C115961682 @default.
- W3188008942 hasConcept C119857082 @default.
- W3188008942 hasConcept C138885662 @default.
- W3188008942 hasConcept C153180895 @default.
- W3188008942 hasConcept C154945302 @default.
- W3188008942 hasConcept C162324750 @default.
- W3188008942 hasConcept C165064840 @default.
- W3188008942 hasConcept C176217482 @default.
- W3188008942 hasConcept C21547014 @default.
- W3188008942 hasConcept C2776401178 @default.
- W3188008942 hasConcept C2781067378 @default.
- W3188008942 hasConcept C33923547 @default.
- W3188008942 hasConcept C41008148 @default.
- W3188008942 hasConcept C41608201 @default.
- W3188008942 hasConcept C41895202 @default.
- W3188008942 hasConcept C52622490 @default.
- W3188008942 hasConcept C83665646 @default.
- W3188008942 hasConceptScore W3188008942C103278499 @default.
- W3188008942 hasConceptScore W3188008942C105795698 @default.
- W3188008942 hasConceptScore W3188008942C108583219 @default.
- W3188008942 hasConceptScore W3188008942C115961682 @default.
- W3188008942 hasConceptScore W3188008942C119857082 @default.
- W3188008942 hasConceptScore W3188008942C138885662 @default.
- W3188008942 hasConceptScore W3188008942C153180895 @default.
- W3188008942 hasConceptScore W3188008942C154945302 @default.
- W3188008942 hasConceptScore W3188008942C162324750 @default.
- W3188008942 hasConceptScore W3188008942C165064840 @default.
- W3188008942 hasConceptScore W3188008942C176217482 @default.
- W3188008942 hasConceptScore W3188008942C21547014 @default.
- W3188008942 hasConceptScore W3188008942C2776401178 @default.
- W3188008942 hasConceptScore W3188008942C2781067378 @default.
- W3188008942 hasConceptScore W3188008942C33923547 @default.
- W3188008942 hasConceptScore W3188008942C41008148 @default.
- W3188008942 hasConceptScore W3188008942C41608201 @default.
- W3188008942 hasConceptScore W3188008942C41895202 @default.
- W3188008942 hasConceptScore W3188008942C52622490 @default.
- W3188008942 hasConceptScore W3188008942C83665646 @default.
- W3188008942 hasFunder F4320321001 @default.
- W3188008942 hasLocation W31880089421 @default.
- W3188008942 hasLocation W31880089422 @default.
- W3188008942 hasOpenAccess W3188008942 @default.
- W3188008942 hasPrimaryLocation W31880089421 @default.