Skip to main content

Table 7 Comparison of accuracy on different datasets (KTH, UCF-11, AVA, collective action)

From: A comparison on visual prediction models for MAMO (multi activity-multi object) recognition using deep learning

Dataset

References

Technique used

Earlier result (%)

Proposed result (%)

KTH

Jhuang et al. [27]

SVM

91.70

98.47

Lin et al. [28]

k-NN

93.43

Liu et al. [29]

Adaboost with C.45

93.80

Kim et al. [30]

NN

95.33

UCF-11

Liu et al. [26]

Adaboost with C.45

71.2

99.52

Cho et al. [15]

SVM and Kernel group sparsity

84.2

Ravanbakhsh et al. [31]

CNN

88.0

89.5

AVA

Ulutan et al. [32]

Actor conditioned attention maps

97.10

99.74

99.74

Collective action

Choi and Savarese [33]

Multiclass SVM

79.2

99.97

Choi et al. [34]

Randomized spatio-temporal volume (RSTV)

82.0