|
|
Using Mutual Information For Model Combination and Automatic Model Switching
We have developed a new algorithm in which the concept of mutual
information is used to combine complementary types of
object models for robust object detection in still images. A
variant of the algorithm automatically switches between multiple
illumination models according to the current context.
|
Model Combination
Combining different and complementary types of object models promises
to increase the robustness and generality of today's computer vision
algorithms. A new method for combining different object models is
introduced which determines a configuration of the models which
maximizes their mutual information. The combination scheme consequently
creates a unified hypothesis from multiple object models on-the-fly
without prior training. To validate the effectiveness of the proposed
method, the approach is applied to the detection of faces combining the
output of three different models in a hierarchy. |
 |
Automatic Model Switching
A major challenge for real-world object tracking is the dynamic nature
of the environmental conditions with respect to illumination, motion,
visibility, etc. For such an environment which may experience drastic
changes at any time, integration of multiple and complementary cues
promises to increase the robustness of visual tracking. Nevertheless,
one has to expect that false positive tracking will occur. In order to
recover from such tracking failure, a new algorithm has been developed
for automatically choosing the object model which best fits the current
context. The algorithm makes use of mutual information to compare
spatial observation densities to a parameterized expectation,
which is specific for the target-object. In order to validate the
effectiveness of the proposed model switching scheme, it is integrated
into a multi-cue face tracking system and experimentally evaluated. |
Publications:
-
Mutual Information For Evidence
Fusion
Hannes Kruppa and Bernt Schiele. In Section "Sensor Fusion,
Registration and Planning" (http://www.dai.ed.ac.uk/CVonline/fusion.htm) of CVonline: On-Line Compendium
of Computer Vision [Online].
R. Fisher (ed).
-
Towards Robust Perception and
Model Integration Bernt Schiele, Martin Spengler,
and Hannes Kruppa. Sensor Based Intelligent
Robots, Lecture Notes in Computer Science, Springer, 2001
-
Hierarchical Combination of Object
Models using Mutual Information Hannes Kruppa and Bernt
Schiele,British Machine Vision Conference, BMVC
2001, Manchester, UK, September 2001
-
Context-driven Model Switching For Visual Tracking
H.Kruppa, B. Schiele. In: 9th
International Symposium on Intelligent Robotic Systems 2001,
Toulouse, France, July 2001
-
Using Mutual Information to Combine Object Models
H.Kruppa, B. Schiele. In 8th International
Symposium on Intelligent Robotic Systems 2000,
Reading, UK, July 2000
Contact:
|
by
webmfritz
—
last modified
2005-10-06 12:14
|
