Transferable Feature Representation for Visible-to-Infrared Cross-Dataset Human Action Recognition

Figure 6

Comparison of visible and infrared actions. The left of each subfigure is the visible image from video sequences in the XD145 dataset, while the right is the infrared image from video sequences in the InfAR dataset.

(a) Fight

(b) Handclapping

(c) Handshake

(d) Hug

(e) Jog

(f) Jump

(g) Punch

(h) Push

(i) Skip

(j) Walk

(k) Wave 1

(l) Wave 2