Definitions - Information Fusion in Signal and Image Processing - page 10

A third model, “orthogonal” to the previous one, consists of combining all of the
M i related to the same decision d i using an operation F , in order to obtain a fused
form M i = F ( M i ,M i ,...,M i ). A decision is then made based on the result of
this combination. In this case, no intermediate decision is made and the information
is handled within the chosen formalism up until the last step, thus reducing contra-
dictions and conflicts. This model, just like the global model, is a centralized model
that requires all of the sources to be available simultaneously. Simpler than the global
model, it is not as flexible as the distributed model, making the possible addition of
sources of information more difficult.
Finally, an intermediate, hybrid model consists of choosing adaptively which infor-
mation is necessary for a given problem based on the specificities of the sources. This
type of model often copies the human expert and involves symbolic knowledge of
the sources and objects. It is therefore often used in rule-based systems. Multi-agent
architectures are well suited for this model.
The system aspect of fusion will be discussed further in an example in Chapter 10.
1.6. Fusion in signal and image processing and fusion in other fields
Fusion in signal and image processing has specific features that need to be taken
into account at every step when constructing a fusion process. These specificities also
require modifying and complexifying certain theoretical tools, often taken from other
fields. This is typically the case of spatial information in image fusion or in robotics.
These specificities will be discussed in detail in the case of fusion in signal, image and
robotics in the following chapters.
The quality of the data to be processed and its heterogenity are often more signif-
icant than in other fields (problems in combining expert opinions, for example). This
causes an additional level of complexity, which has to be taken into account in the
modeling, but also in the algorithms.
The data is mostly objective (provided by sensors), which separates them from
subjective data such as what can be provided by individuals. However, they maintain
a certain part of subjectivity (for example, in the choice of the sensors or the sources
of information, or also of the acquisition parameters). There is also some subjectivity
in how the objectives are expressed. Objective data is usually degraded, either because
of imperfection in the acquisition systems, or because of the processes to which it is
subjected.
In fact, one of the main difficulties comes from the fact that the types of knowledge
that are dealt with are very heterogenous. They are comprised not just of measure-
ments and observations (which can be heterogenous themselves), but also of general
cases, typical examples, generic models, etc.