Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313291202> ?p ?o ?g. }
- W4313291202 endingPage "3397" @default.
- W4313291202 startingPage "3383" @default.
- W4313291202 abstract "Group activity recognition, which aims to simultaneously understand individual action and group activity in video clips, plays a fundamental role in computer vision and video analysis. In this paper, we propose a novel relational inference framework, termed MLST-Former, for individual action and group activity recognition, capturing the various degrees of spatial-temporal dynamic interactions adaptively and jointly among actors and generating reasonable group representations. Specifically, we first design a multi-level spatial-temporal Transformer to capture the miscellaneous actors’ spatial-temporal contextual information to deal with unbalanced interactions between actors. Furthermore, our proposed network is capable of fully mining long-range spatial-temporal dependencies with the virtue of the merge function and cross attention mechanism. We then propose an inter-frame gating fusion mechanism (IGFM) to selectively aggregate the temporal and structural features of the interacting actors. A new multi-task learning strategy, consisting of the classification cost of individual actions and group activities, is also developed. Moreover, we adopt the motion trajectory branch to provide complementary dynamic features for improving recognition performance. A series of ablation studies demonstrate the effectiveness and respective contributions of the different components within the proposed method. Extensive experiments on four public GAR datasets clearly show that our approach can achieve very competitive performance by comparing them with state-of-the-art methods." @default.
- W4313291202 created "2023-01-06" @default.
- W4313291202 creator A5002051996 @default.
- W4313291202 creator A5039458023 @default.
- W4313291202 creator A5047508906 @default.
- W4313291202 creator A5062373599 @default.
- W4313291202 creator A5087818121 @default.
- W4313291202 date "2023-07-01" @default.
- W4313291202 modified "2023-10-09" @default.
- W4313291202 title "MLST-Former: Multi-Level Spatial-Temporal Transformer for Group Activity Recognition" @default.
- W4313291202 cites W100367037 @default.
- W4313291202 cites W179875071 @default.
- W4313291202 cites W1969752788 @default.
- W4313291202 cites W2002741386 @default.
- W4313291202 cites W2047499569 @default.
- W4313291202 cites W2064675550 @default.
- W4313291202 cites W2081314772 @default.
- W4313291202 cites W2100304967 @default.
- W4313291202 cites W2110275531 @default.
- W4313291202 cites W2134282114 @default.
- W4313291202 cites W2171544105 @default.
- W4313291202 cites W2183341477 @default.
- W4313291202 cites W2259801182 @default.
- W4313291202 cites W2269938945 @default.
- W4313291202 cites W2552950674 @default.
- W4313291202 cites W2733111742 @default.
- W4313291202 cites W2736442062 @default.
- W4313291202 cites W2778252923 @default.
- W4313291202 cites W2891433740 @default.
- W4313291202 cites W2892664578 @default.
- W4313291202 cites W2897054934 @default.
- W4313291202 cites W2914868535 @default.
- W4313291202 cites W2920457288 @default.
- W4313291202 cites W2927778007 @default.
- W4313291202 cites W2940963663 @default.
- W4313291202 cites W2947315758 @default.
- W4313291202 cites W2961553857 @default.
- W4313291202 cites W2963047744 @default.
- W4313291202 cites W2963076818 @default.
- W4313291202 cites W2963377215 @default.
- W4313291202 cites W2963524571 @default.
- W4313291202 cites W2964062686 @default.
- W4313291202 cites W2964298947 @default.
- W4313291202 cites W2974686944 @default.
- W4313291202 cites W3006198072 @default.
- W4313291202 cites W3014545861 @default.
- W4313291202 cites W3034934229 @default.
- W4313291202 cites W3035029089 @default.
- W4313291202 cites W3035717782 @default.
- W4313291202 cites W3095022195 @default.
- W4313291202 cites W3095422700 @default.
- W4313291202 cites W3108614371 @default.
- W4313291202 cites W3119381934 @default.
- W4313291202 cites W3119771537 @default.
- W4313291202 cites W3121052081 @default.
- W4313291202 cites W3126367810 @default.
- W4313291202 cites W3160569599 @default.
- W4313291202 cites W3162363705 @default.
- W4313291202 cites W3169998028 @default.
- W4313291202 cites W3171529719 @default.
- W4313291202 cites W3175859725 @default.
- W4313291202 cites W3186621246 @default.
- W4313291202 cites W3204485253 @default.
- W4313291202 cites W4205432963 @default.
- W4313291202 cites W4206654595 @default.
- W4313291202 cites W4214759957 @default.
- W4313291202 cites W4221060219 @default.
- W4313291202 cites W4230451144 @default.
- W4313291202 cites W4287734800 @default.
- W4313291202 cites W4289822576 @default.
- W4313291202 cites W4304080324 @default.
- W4313291202 cites W4312349914 @default.
- W4313291202 cites W4312375685 @default.
- W4313291202 cites W4312557989 @default.
- W4313291202 cites W4312919330 @default.
- W4313291202 doi "https://doi.org/10.1109/tcsvt.2022.3233069" @default.
- W4313291202 hasPublicationYear "2023" @default.
- W4313291202 type Work @default.
- W4313291202 citedByCount "1" @default.
- W4313291202 crossrefType "journal-article" @default.
- W4313291202 hasAuthorship W4313291202A5002051996 @default.
- W4313291202 hasAuthorship W4313291202A5039458023 @default.
- W4313291202 hasAuthorship W4313291202A5047508906 @default.
- W4313291202 hasAuthorship W4313291202A5062373599 @default.
- W4313291202 hasAuthorship W4313291202A5087818121 @default.
- W4313291202 hasConcept C119857082 @default.
- W4313291202 hasConcept C121687571 @default.
- W4313291202 hasConcept C124101348 @default.
- W4313291202 hasConcept C153180895 @default.
- W4313291202 hasConcept C154945302 @default.
- W4313291202 hasConcept C197129107 @default.
- W4313291202 hasConcept C23123220 @default.
- W4313291202 hasConcept C2776214188 @default.
- W4313291202 hasConcept C2777212361 @default.
- W4313291202 hasConcept C2987834672 @default.
- W4313291202 hasConcept C41008148 @default.
- W4313291202 hasConceptScore W4313291202C119857082 @default.
- W4313291202 hasConceptScore W4313291202C121687571 @default.