State-of-the-Art

Toyota Smarthome Trimmed

Here we report the mean per-class accuracy. Note that, we provide two versions of Poses: Pose_V1.1 are the poses extracted by LCRNET. Pose_V1.2 are the poses further refined by SSTA-PRS. Last updated: 01/02/2021.

Method Modality CS CV1 CV2
DT RGB 41.9 20.9 23.7
LSTM Pose_V1.1 42.5 13.4 17.2
I3D RGB 53.4 34.9 45.1
I3D+NL RGB 53.6 34.3 43.9
Separable-STA RGB+Pose_V1.1 54.2 35.2 50.3
5C-AGCN+PRS Pose_V1.2 62.1 22.8 54.0
VPN RGB+Pose_V1.1 60.8 43.8 53.5
VPN RGB+Pose_V1.2 65.2 54.1
AssembleNet++ RGB 63.6

Toyota Smarthome Untrimmed

Here we report the frame-based mAP. The Poses are extracted based on LCRNET+SSTA-PRS. Last updated: 01/02/2021.

Method Modality CS CV
AGCN+LSTM Pose 17.0 14.8
I3D+LSTM RGB 22.6 12.9
I3D+Dilated-TCN RGB 25.1 13.9
I3D+MS-TCN RGB 25.9 13.1
I3D+TGM RGB 26.7 13.4
I3D+PDAN RGB 32.7

Comments are closed.

  • Bibtex of Toyota Smarthome Trimmed Dataset:

    @InProceedings{Das_2019_ICCV,
        author = {Das, Srijan and Dai, Rui and Koperski, Michal and Minciullo, Luca and Garattoni, Lorenzo and Bremond, Francois and Francesca, Gianpiero},
        title = {Toyota Smarthome: Real-World Activities of Daily Living},
        booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
        month = {October},
        year = {2019}
    }
  • Bibtex of Toyota Smarthome Untrimmed Dataset:

    @misc{dai2020toyota,
        author = {Dai, Rui and Das, Srijan and Sharma, Saurav and Minciullo, Luca and Garattoni, Lorenzo and Bremond, Francois and Francesca, Gianpiero},
        title = {Toyota Smarthome Untrimmed: Real-World Untrimmed Videos for Activity Detection}, 
        year = {2020}, 
        eprint = {2010.14982}, 
        archivePrefix = {arXiv}, 
        primaryClass = {cs.CV}
    }