Computational Intelligence and Neuroscience

Research Article

Multiview Layer Fusion Model for Action Recognition Using RGBD Images

Table 2

Sampled features in single view and fused features in multiview.


Action/view	Camera 1	Camera 2
Action/view	I_mo	I_mo

(I) Standing-walking/front
(II) Standing-walking/slant
(III) Standing-walking/side
(IV) Standing-walking/back slant
(V) Standing-walking/back
(VI) Sitting/front
(VII) Sitting/slant
(VIII) Sitting/side
(IX) Sitting/back slant
(X) Sitting/back
(XI) Stooping/front
(XII) Stooping/slant
(XIII) Stooping/side
(XIV) Stooping/back slant
(XV) Stooping/back
(XVI) Lying/front
(XVII) Lying/slant
(XVIII) Lying/side
(XIX) Lying/back slant
(XX) Lying/back

Note: (I) is weighted depth density of layers which uses α for adjustment between and . (II) , fused features from multiview, called the weighted mass of dimension feature.