Database Reference
In-Depth Information
Fig. 11.2
Illustration of VR
dance training with feedback
protocols, including (
a
)
side-by-side feedback, (
b
)
overlay feedback, and (
c
)
score-curve feedback
frame content. The SSOM parser is performed on each frame of the gesture, thus
creating the coding symbols showing a trajectory on the SSOM. The postures are
characterized by a template calculated by the bag-of-word (BoW) model [
350
], the
sparse code [
351
,
352
], or a newly proposed method based on the posture transition
matrix. For gesture recognition, the template obtained from the input gesture is
compared with a set of reference templates using a matching algorithm that provides
Search WWH ::
Custom Search