Information Technology Reference
In-Depth Information
(a)
(b)
(c)
(d)
Fig. 9 Examples of human detection in the ticket office with heavy background clutter. (a) the
original slice; (b) initial hypothesises use pure static human detector; (c) motion confidence
map; (d) detection results using the motion map of (c).
(a)
(b)
Fig. 10 Gaussian weighting masks corresponding to the hypothesises bounding box. (a) cor-
responds to Fig. 8(e); (b) corresponds to Fig. 9(d).
4
Discussion and Conclusions
In this paper, we presented a framework for robust people detection in highly clut-
tered scenes with low resolution image sequences. In particular, we introduced a
Search WWH ::




Custom Search