|
View-invariant Recognition of Human Actions
Recognition of human motion and action is an important
area of research in computer vision that plays a crucial
role in various applications such as surveillance, humancomputer
interaction, ergonomics, etc. Depending on the
nature and the scale of the problem, understanding of human
motion is usually studied at four levels of granularity:
pose, movement, action, and activity. The
challenges being addressed include perspective distortions,
differences in viewpoints, unknown camera parameters, anthropometric
variations, and the large degrees of freedom
of articulated bodies. Action can be regarded as a
collection of 4D space-time data observed by a perspective
video camera. After image projection, the 3D Euclidean
information is lost and projectively distorted, which makes
action recognition rather challenging, especially for varying
viewpoints. Another source of challenge is the irregularities
of human actions due to a variety of factors such as age,
gender, circumstances, etc. The timeline of action is another
important issue in action recognition. The execution
rates of the same action in different videos may vary due
to different actors or variable camera frame rates. Therefore,
the mapping between same actions in different videos
is usually highly non-linear.
More details can be found HERE.
Related papers:
View-invariant Recognition of Human Body Poses
More details can be found HERE.
Related papers:
Video Completion / Editing
Video completion is of great importance to many applications such as
video repairing, video editing and movie post-production. We propose a
novel technique to fill in missing
background and moving foreground of a video captured by
a static or moving camera. Different from previous work
which are based on processing in the 3D data volume, we
slice the volume along the motion manifold of the moving
object, and therefore reduce the search space from 3D
to 2D, while still preserving the spatial smoothness and
temporal coherence. In addition to the computational efficiency,
based on geometric video analysis, the proposed
approach is also able to handle real videos under perspective
distortion, as well as common camera motions, such
as panning, translation and zooming. The experimental results
demonstrate that our algorithm performs comparably
to 3D search based methods, while extending the current
state-of-the-art repairing techniques to videos with significant
camera motion and projective effects, as well as lighting
changes.
More results can be found HERE.
Related papers:
Shadow Composition
We suggested a novel approach to create visually realistic and
geometrically correct shadows for objects that are composited
into a single view of a target scene.
Compared to traditional single view
compositing methods, which either
do not deal with the shadow effects
or manually create the shadows for
the composited objects, our approach
efficiently utilizes the geometric and
photometric constraints extracted
from a single target image to synthesize
the shadows consistent with the
overall target scene for the inserted
objects.
Related papers:
|