Matches in SemOpenAlex for { <https://semopenalex.org/work/W3173567920> ?p ?o ?g. }
- W3173567920 abstract "With the rapid development of social media, tremendous videos with new classes are generated daily, which raise an urgent demand for video classification methods that can continuously update new classes while maintaining the knowledge of old videos with limited storage and computing resources. In this paper, we summarize this task as Class-Incremental Video Classification (CIVC) and propose a novel framework to address it. As a subarea of incremental learning tasks, the challenge of catastrophic forgetting is unavoidable in CIVC. To better alleviate it, we utilize some characteristics of videos. First, we decompose the spatio-temporal knowledge before distillation rather than treating it as a whole in the knowledge transfer process; trajectory is also used to refine the decomposition. Second, we propose a dual granularity exemplar selection method to select and store representative video instances of old classes and key-frames inside videos under a tight storage budget. We benchmark our method and previous SOTA class-incremental learning methods on Something-Something V2 and Kinetics datasets, and our method outperforms previous methods significantly." @default.
- W3173567920 created "2021-07-05" @default.
- W3173567920 creator A5007599551 @default.
- W3173567920 creator A5007983738 @default.
- W3173567920 creator A5027826355 @default.
- W3173567920 creator A5056078581 @default.
- W3173567920 creator A5087698461 @default.
- W3173567920 date "2021-06-30" @default.
- W3173567920 modified "2023-09-27" @default.
- W3173567920 title "When Video Classification Meets Incremental Classes" @default.
- W3173567920 cites W114517082 @default.
- W3173567920 cites W136504859 @default.
- W3173567920 cites W1522734439 @default.
- W3173567920 cites W1682403713 @default.
- W3173567920 cites W1821462560 @default.
- W3173567920 cites W1920196880 @default.
- W3173567920 cites W1965115694 @default.
- W3173567920 cites W1983364832 @default.
- W3173567920 cites W1997865285 @default.
- W3173567920 cites W2105101328 @default.
- W3173567920 cites W2108598243 @default.
- W3173567920 cites W2126574503 @default.
- W3173567920 cites W2156303437 @default.
- W3173567920 cites W2342662179 @default.
- W3173567920 cites W2473930607 @default.
- W3173567920 cites W2507009361 @default.
- W3173567920 cites W2526479943 @default.
- W3173567920 cites W2553594924 @default.
- W3173567920 cites W2618767506 @default.
- W3173567920 cites W2619947201 @default.
- W3173567920 cites W2625366777 @default.
- W3173567920 cites W2736596806 @default.
- W3173567920 cites W2746726611 @default.
- W3173567920 cites W2765895026 @default.
- W3173567920 cites W2770804203 @default.
- W3173567920 cites W2786446225 @default.
- W3173567920 cites W2790486454 @default.
- W3173567920 cites W2798472916 @default.
- W3173567920 cites W2799176631 @default.
- W3173567920 cites W2883429621 @default.
- W3173567920 cites W2884282566 @default.
- W3173567920 cites W2891446678 @default.
- W3173567920 cites W2899771611 @default.
- W3173567920 cites W2934628279 @default.
- W3173567920 cites W2948734064 @default.
- W3173567920 cites W2954929116 @default.
- W3173567920 cites W2962934715 @default.
- W3173567920 cites W2963155035 @default.
- W3173567920 cites W2963524571 @default.
- W3173567920 cites W2963645879 @default.
- W3173567920 cites W2963820951 @default.
- W3173567920 cites W2964189064 @default.
- W3173567920 cites W2981385151 @default.
- W3173567920 cites W2984287396 @default.
- W3173567920 cites W2990152177 @default.
- W3173567920 cites W2990503944 @default.
- W3173567920 cites W2996901793 @default.
- W3173567920 cites W2996937600 @default.
- W3173567920 cites W2997563949 @default.
- W3173567920 cites W3030364939 @default.
- W3173567920 cites W3034572008 @default.
- W3173567920 cites W3035303837 @default.
- W3173567920 cites W3035374961 @default.
- W3173567920 cites W3035501943 @default.
- W3173567920 cites W3036967712 @default.
- W3173567920 cites W3094953068 @default.
- W3173567920 cites W3096833468 @default.
- W3173567920 cites W3100156920 @default.
- W3173567920 cites W3105913416 @default.
- W3173567920 cites W3107810305 @default.
- W3173567920 cites W3113584014 @default.
- W3173567920 cites W3133570661 @default.
- W3173567920 cites W3157898411 @default.
- W3173567920 doi "https://doi.org/10.48550/arxiv.2106.15827" @default.
- W3173567920 hasPublicationYear "2021" @default.
- W3173567920 type Work @default.
- W3173567920 sameAs 3173567920 @default.
- W3173567920 citedByCount "0" @default.
- W3173567920 crossrefType "posted-content" @default.
- W3173567920 hasAuthorship W3173567920A5007599551 @default.
- W3173567920 hasAuthorship W3173567920A5007983738 @default.
- W3173567920 hasAuthorship W3173567920A5027826355 @default.
- W3173567920 hasAuthorship W3173567920A5056078581 @default.
- W3173567920 hasAuthorship W3173567920A5087698461 @default.
- W3173567920 hasBestOaLocation W31735679201 @default.
- W3173567920 hasConcept C111919701 @default.
- W3173567920 hasConcept C119857082 @default.
- W3173567920 hasConcept C124952713 @default.
- W3173567920 hasConcept C13280743 @default.
- W3173567920 hasConcept C138885662 @default.
- W3173567920 hasConcept C142362112 @default.
- W3173567920 hasConcept C154945302 @default.
- W3173567920 hasConcept C162324750 @default.
- W3173567920 hasConcept C177774035 @default.
- W3173567920 hasConcept C185798385 @default.
- W3173567920 hasConcept C187736073 @default.
- W3173567920 hasConcept C205649164 @default.
- W3173567920 hasConcept C26517878 @default.
- W3173567920 hasConcept C2777212361 @default.
- W3173567920 hasConcept C2780451532 @default.