Human Interaction Recognition on UT benchmark leaderboard · SOTA2 Research

Human Interaction Recognition on UT

95Average AccuracyZhang et al. [56]

Evaluation Results

Method	Links
Zhang et al. [56]2017.06		95	100	100	100	90	90	90
Wang et al. [46]2017.06		95	—	—	—	—	—	—
Co-LSTSM2017.06		95	100	100	90	100	90	90
Kong et al. [16]2017.06		93.33	90	100	90	100	90	90
Ke et al. [15]2017.06		93.33	—	—	—	—	—	—
Raptis & Sigal [31]2017.06		93.3	100	100	90	100	80	90
Kong et al. [19]2017.06		91.67	100	90	100	80	90	90
Shariat & Pavlovic [36]2017.06		91.57	—	—	—	—	—	—
Two CNN+LSTM2017.06		90	100	100	90	80	90	80
Kong et al. [18]2017.06		88.33	80	80	100	90	90	90
Ryoo [32]2017.06		85	80	90	90	80	90	80
Donahue et al. [8]2017.06		85	90	80	90	80	90	80
One CNN+LSTM2017.06		85	90	80	90	80	90	80
Yu et al. [53]2017.06		83.33	100	65	100	85	75	75
Person-box CNN2017.06		81.67	90	80	80	80	80	80
Ryoo et al. [34]2017.06		70.8	75	87.5	62.5	50	75	75