Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202364407> ?p ?o ?g. }
- W3202364407 endingPage "4169" @default.
- W3202364407 startingPage "4158" @default.
- W3202364407 abstract "Person Re-identification (ReID) aims at matching a person of interest across images. In convolutional neural network (CNN) based approaches, loss design plays a vital role in pulling closer features of the same identity and pushing far apart features of different identities. In recent years, triplet loss achieves superior performance and is predominant in ReID. However, triplet loss considers only three instances of two classes in per-query optimization (with an anchor sample as query) and it is actually equivalent to a two-class classification. There is a lack of loss design which enables the joint optimization of multiple instances (of multiple classes) within per-query optimization for person ReID. In this paper, we introduce a multi-class classification loss, <italic xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>i</i> . <italic xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>e</i> ., N-tuple loss, to jointly consider multiple ( <inline-formula xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink><tex-math notation=LaTeX>$N$</tex-math></inline-formula> ) instances for per-query optimization. This in fact aligns better with the ReID test/inference process, which conducts the ranking/comparisons among multiple instances. Furthermore, for more efficient multi-class classification, we propose a new meta prototypical N-tuple loss. With the multi-class classification incorporated, our model achieves the state-of-the-art performance on the benchmark person ReID datasets" @default.
- W3202364407 created "2021-10-11" @default.
- W3202364407 creator A5008582408 @default.
- W3202364407 creator A5049237194 @default.
- W3202364407 creator A5049963367 @default.
- W3202364407 creator A5051469345 @default.
- W3202364407 creator A5079572598 @default.
- W3202364407 date "2022-01-01" @default.
- W3202364407 modified "2023-10-16" @default.
- W3202364407 title "Beyond Triplet Loss: Meta Prototypical N-Tuple Loss for Person Re-identification" @default.
- W3202364407 cites W1920259731 @default.
- W3202364407 cites W1928419358 @default.
- W3202364407 cites W1982925187 @default.
- W3202364407 cites W2108598243 @default.
- W3202364407 cites W2138621090 @default.
- W3202364407 cites W2194775991 @default.
- W3202364407 cites W2201508557 @default.
- W3202364407 cites W2204750386 @default.
- W3202364407 cites W2467139031 @default.
- W3202364407 cites W2471048925 @default.
- W3202364407 cites W2491664569 @default.
- W3202364407 cites W2499468060 @default.
- W3202364407 cites W2502225121 @default.
- W3202364407 cites W2520774990 @default.
- W3202364407 cites W2584637367 @default.
- W3202364407 cites W2585635281 @default.
- W3202364407 cites W2596603442 @default.
- W3202364407 cites W2606377603 @default.
- W3202364407 cites W2618530766 @default.
- W3202364407 cites W2625674597 @default.
- W3202364407 cites W2724213014 @default.
- W3202364407 cites W2736410039 @default.
- W3202364407 cites W2738760914 @default.
- W3202364407 cites W2752782242 @default.
- W3202364407 cites W2755066373 @default.
- W3202364407 cites W2777534232 @default.
- W3202364407 cites W2798775284 @default.
- W3202364407 cites W2798794112 @default.
- W3202364407 cites W2883348239 @default.
- W3202364407 cites W2884366600 @default.
- W3202364407 cites W2896888563 @default.
- W3202364407 cites W2962926870 @default.
- W3202364407 cites W2963047834 @default.
- W3202364407 cites W2963049565 @default.
- W3202364407 cites W2963078173 @default.
- W3202364407 cites W2963438548 @default.
- W3202364407 cites W2963664762 @default.
- W3202364407 cites W2963805953 @default.
- W3202364407 cites W2963842104 @default.
- W3202364407 cites W2963901085 @default.
- W3202364407 cites W2964289004 @default.
- W3202364407 cites W2966961956 @default.
- W3202364407 cites W2971765950 @default.
- W3202364407 cites W2980073905 @default.
- W3202364407 cites W2981420411 @default.
- W3202364407 cites W2984145721 @default.
- W3202364407 cites W2994675267 @default.
- W3202364407 cites W2997987796 @default.
- W3202364407 cites W2998508940 @default.
- W3202364407 cites W3000133216 @default.
- W3202364407 cites W3005339439 @default.
- W3202364407 cites W3034580371 @default.
- W3202364407 cites W3035724178 @default.
- W3202364407 cites W3097870364 @default.
- W3202364407 cites W3099206234 @default.
- W3202364407 cites W3100506510 @default.
- W3202364407 cites W3101227480 @default.
- W3202364407 cites W3112897560 @default.
- W3202364407 doi "https://doi.org/10.1109/tmm.2021.3115451" @default.
- W3202364407 hasPublicationYear "2022" @default.
- W3202364407 type Work @default.
- W3202364407 sameAs 3202364407 @default.
- W3202364407 citedByCount "2" @default.
- W3202364407 countsByYear W32023644072023 @default.
- W3202364407 crossrefType "journal-article" @default.
- W3202364407 hasAuthorship W3202364407A5008582408 @default.
- W3202364407 hasAuthorship W3202364407A5049237194 @default.
- W3202364407 hasAuthorship W3202364407A5049963367 @default.
- W3202364407 hasAuthorship W3202364407A5051469345 @default.
- W3202364407 hasAuthorship W3202364407A5079572598 @default.
- W3202364407 hasBestOaLocation W32023644072 @default.
- W3202364407 hasConcept C105795698 @default.
- W3202364407 hasConcept C116834253 @default.
- W3202364407 hasConcept C118615104 @default.
- W3202364407 hasConcept C118930307 @default.
- W3202364407 hasConcept C119857082 @default.
- W3202364407 hasConcept C13280743 @default.
- W3202364407 hasConcept C154945302 @default.
- W3202364407 hasConcept C165064840 @default.
- W3202364407 hasConcept C185798385 @default.
- W3202364407 hasConcept C189430467 @default.
- W3202364407 hasConcept C205649164 @default.
- W3202364407 hasConcept C23123220 @default.
- W3202364407 hasConcept C2777212361 @default.
- W3202364407 hasConcept C33923547 @default.
- W3202364407 hasConcept C41008148 @default.
- W3202364407 hasConcept C59822182 @default.
- W3202364407 hasConcept C80444323 @default.