16-09-2014, 12:03 PM
Upper Body Human Detection and Segmentation in
Low Contrast Video
Upper Body.pdf (Size: 2.23 MB / Downloads: 186)
Abstract—
In the application of extracting human regions from
videos, many existing methods may lose their efficacy when
illumination varies or the human remains still. To address this
problem, we propose a method in this paper for human region
detection and segmentation by constructing a generalized human
upper body model. The method mainly consists of two main
procedures. Firstly, foreground connected regions are extracted
by background subtraction from current frame and classified
through a human upper body model pre-trained with SVM
(Support Vector Machine) to determine whether they are human
regions. Secondly, we assign an energy function to the region
contour and apply an energy minimization procedure to evolve
the contour when human regions are “polluted” by background,
for example, changes of lighting conditions. After finding the
optimal contour we update the background and repeat the
procedures in next frame. This feedback strategy rectifies the
mistaken background regions promptly and extracts human
regions correctly. Our experimental results demonstrate that
the proposed method is robust enough to handle videos of low
contrast as well as normal conditions.
B. Segmentation
Foreground segmentation, or background subtraction, has a
longer history. Stauffer et al. [16] represented each pixel by
a mixture of Gaussian distributions and updated each pixel
with new Gaussians. Kim et al. [17] expressed background as
cylindrical codebook model and updated the model adaptively.
Sample background values at each pixel are quantized into
codebooks which represent a compressed form of background
model for a long image sequence. Liu et al. [18] sampled
pixels along the time axis and clustered them with mean
shift technique. Each cluster was assigned with a weight to
demonstrate its likelihood of being background. Though these
methods can handle illumination variation, they still fail once
foreground objects keep still for a long time
CONCLUSION
Our proposed method cannot be used on static images and
currently it can only detect one object. This is a limitation of
our method.
ACKN