| Presentation |
| Type of Publication |
| Video-Object Segmentation: from MPEG-4 Coding to Behaviour
Analysis |
| Title |
|
| Authors |
| Presentation,
Zoran corp., Santa Clara, 02. May 2003 |
| Published in |
| Current video-processing systems still treat
video as rectangular images without further structure. However,
humans watching the video immediately recognize acting objects as
semantic units. This semantic object separation is currently not
reflected at the technical side, making it difficult to
manipulate and analyse the video at the object level. Enabling
object based manipulation will introduce many new possibilities
for working with videos like composing new scenes from
pre-existing video-objects, or providing the possibility for
user-interaction with the scene. Moreover, object segmentation is
required for a classification of video-objects, content-based
retrieval, and video analysis. A central problem is how to bridge
the gap between low-level video processing on the signal level to
semantically meaningful high-level objects. It seems that a
strictly multi-stage process, which provides increasingly
semantically meaningful data is not a good approach since errors
at lower levels are hard to correct at later stages. To achieve a
good segmentation, it is required to integrate an object-model
into early stages of the the segmentation process. This model
carries information about the high-level object structure and
helps the segmentation algorithm to make the right decisions at
low processing levels.This presentation outlines a segmentation
algorithm which is based on camera-motion compensationed
background subtraction to detect foreground objects. The
algorithm is designed to provide all the information for an
MPEG-4 encoder utilizing shape encoding and background-sprites.
It is further shown, how object models can be integrated to
improve the semantic accuracy and how video-object behaviour can
be extracted based on the obtained segmentation masks. |
| Abstract |
|
MPEG-4
segmentation
parametric motion-estimation
background reconstruction
object model
|
| Keywords |
| [PDF]
[BIB] [XML] |
| Downloads & Bib-Entries |
