Skip to main content
. 2017 Mar 22;11:13. doi: 10.3389/fnbot.2017.00013

Table 4.

Weizmann comparison to other approaches.

Catagory # Actions Evaluation fps Temporal range
Sub-sequence Full sequence
Recall # Frames Recall
Key pose Weinland and Boyer, 2008 10 LOSO 93.6
Baysal et al., 2010 9 LOO 92.6
Cheema et al., 2011 9 LOO 91.6
Chaaraoui et al., 2013 9 LOSO 124 92.8
Liu et al., 2013 10 LOSO 100
Single frame Niebles and Fei-Fei, 2007 9 LOSO 55 1 72.8
Fathi and Mori, 2008 10 LOO 0.25–5 99.9 1 100
Schindler and van Gool, 2008 9 LOSO 93.5 1 100
Hoai et al., 2011 10 LOSO 87.7 1
Full sequence Jhuang et al., 2007 9 0.83 98.8
Klaser et al., 2008 10 LOSO 84.3
Grundmann et al., 2008 9 LOAO 94.6
Ikizler and Duygulu, 2009 9 LOO 100
Bregonzio et al., 2009 10 LOSO 96.7
Sun and Liu, 2012 10 LOO 97.8
Beaudry et al., 2016 10 LOO 51.63 100
Presented approach 10 LOSO 1, 000 82.2 1 92.2

Bold values indicate maximum recall/fps values per column.