A curated list of Group Activity (a.k.a., Collective Activity) recognition papers, codes, datasets and other resources. If you have any problems, suggestions or improvements, please submit an issue or PR.
- Paper list
- Available codes
- Leaderboard
- (2016) Volleyball Dataset [Github]
- (2012) Choi's Dataset [Homepage]
- (2011) Collective Activity Augmented Dataset [Homepage]
- (2009) Collective Activity Dataset (CAD) [Homepage]
This section only shows some popular or new datasets.
This section only includes the last five papers since 2018 in arXiv.org. Note that arXiv papers without available codes are not included in the leaderboard of performance.
- A Multi-Stream Convolutional Neural Network Framework for Group Activity Recognition, 2018.12 [arxiv]
- GAIM: Graph Attention Interaction Model for Collective Activity Recognition (T-MM 2020) [paper]
- CCG-LSTM: Coherence Constrained Graph LSTM for Group Activity Recognition (T-PAMI 2019) [paper]
- stagNet: An Attentive Semantic RNN for Group Activity and Individual Action Recognition (TCSVT 2019) [paper]
- GCN+SPA+KD+OF: Learning Semantics-Preserving Attention and Contextual Interaction for Group Activity Recognition (TIP 2019) [paper]
- PRL: Progressive Relation Learning for Group Activity Recognition (CVPR 2020) [paper]
- ARG: Learning Actor Relation Graphs for Group Activity Recognition (CVPR 2019) [paper] [github]
- CRM: Convolutional Relational Machine for Group Activity Recognition (CVPR 2019) [paper]
- PC-TDM: Participation-Contributed Temporal Dynamic Model for Group Activity Recognition (ACM MM 2018) [paper] [github]
- SPA+KD+OF: Mining Semantics-Preserving Attention for Group Activity Recognition (ACM MM 2018) [paper]
- RCRG: Hierarchical Relational Networks for Group Activity Recognition and Retrieval (ECCV 2018) [paper] [github]
- stagNet: An Attentive Semantic RNN for Group Activity Recognition (ECCV 2018) [paper]
- HANs+HCNs: Hierarchical Attention and Context Modeling for Group Activity Recognition (ICASSP 2018) [paper]
- SRNN: Structural Recurrent Neural Network (SRNN) for Group Activity Analysis (WACV 2018) [paper]
- MLS-GAN: A Hierarchical Deep Temporal Model for Group Activity Recognition (ACCV 2018) [paper]
- SBGAR: Semantics Based Group Activity Recognition (ICCV 2017) [paper]
- SSU: Social Scene Understanding: End-to-end Multi-person Action Localization and Collective Activity Recognition (CVPR 2017) [paper] [github]
- RMIC: Recurrent Modeling of Interaction Context for Collective Activity Recognition (CVPR 2017) [paper]
- CERN: Confidence-energy Recurrent Network for Group Activity Recognition (CVPR 2017) [paper]
- HDTM: A Hierarchical Deep Temporal Model for Group Activity Recognition (CVPR 2016) [paper] Deep Temporal Model for Group Activity Recognition (CVPR 2016) [paper] [github]
- SIM: Structure Inference Machines: Recurrent Neural Networks for Analyzing Relations in Group Activity Recognition (CVPR 2016) [paper]
- GM: A Generative Model for Recognizing Mixed Group Activities in Still Images (IJCAI 2016) [paper]
- omitted
The section is being continually updated. We only show results on Volleyball and CAD datasets.
MCA: Multi-class Classifcation Accuracy
MPCA: Mean Per Class Accuracy
| Year | Methods | MCA | MPCA |
|---|---|---|---|
| 2016 | HDTM | 81.9 | 82.9 |
| 2017 | SBGAR | 66.9 | 67.6 |
| 2017 | CERN-2 | 83.3 | 83.6 |
| 2017 | SSU | 89.9 | - |
| 2018 | SRNN | 83.5 | - |
| 2018 | HANs+HCNs | 85.1 | - |
| 2018 | PC-TDM | 87.7 | 88.1 |
| 2018 | stagNet | 89.3 | - |
| 2018 | RCRG | 89.5 | - |
| 2018 | SPA+KD+OF | 90.7 | 90.0 |
| 2019 | GCN+SPA+KD+OF | 91.2 | 91.4 |
| 2019 | ARG | 92.6 | - |
| 2019 | CRM | 93.0 | - |
| 2020 | PRL | 91.4 | 91.8 |
| Year | Methods | MCA (5 classes) | MPCA (4 classes) |
|---|---|---|---|
| 2016 | GM | - | 88.9 |
| 2016 | HDTM | 81.5 | 89.6 |
| 2017 | CERN-2 | 87.2 | 88.3 |
| 2017 | RMIC | - | 89.4 |
| 2017 | SBGAR | 86.1 | 89.9 |
| 2018 | HANs+HCNs | 84.3 | - |
| 2018 | PC-TDM | - | 92.2 |
| 2018 | stagNet | 89.1 | - |
| 2018 | SPA+KD+OF | - | 95.7 |
| 2019 | GCN+SPA+KD+OF | - | 95.8 |
| 2019 | CRM | 85.8 | 94.2 |
| 2019 | ARG | 91.0 | - |
| 2020 | PRL | - | 93.8 |