TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
classification layer which classifies the object between them. As we said, the result of the<br />
classification is a set of AVOs sorted according to their importance in the scene. The classification<br />
layer affects a final relative priority score (RPS) to each Access Unit to apply a differentiation<br />
mechanism. This priority score reflects both the priority of particular AVO in the scene and the<br />
priority of a single frame type (I, P, B or hierarchical stream if any BL or EL). In the rest of this<br />
Section, we will interest to the prototype implementation which is based on a neural network<br />
algorithm.<br />
n<br />
Let T = {(X,c)} be a training set of N labeled vectors, where Χ ∈R is a feature vector X =<br />
(x 1 ,x 2 ,…,x n ) t and c ∈ T is its class label from an index set T (X is an AVO, and T is the set of<br />
available classes (e.g. Diffserv PHB), in our context) . The variables x i are referred to as attributes<br />
(e.g. QoS features of each AVO). A class is modeled by one or more prototype, which have<br />
{ U , U , 2<br />
K }<br />
U m<br />
1<br />
as features.<br />
A classifier is a mapping function called C defined as C: R n T, which assigns a class label in<br />
T to each vector in R n (a vector is an MPEG-4 AVO features). Typically, the classifier is<br />
represented by a set of model parameters Λ = { ∆ k<br />
}. The classifier specifies a partitioning of the<br />
feature space into regions R j = { X∈ R n n<br />
: C(X) = j} where U R<br />
j<br />
≡ R and I R<br />
j<br />
≡ Φ .<br />
It also induces a corresponding partitioning of the training set into subset T.<br />
There are different methods based on this definition, which allow an automatic classification<br />
of Vectors. We present in what below the classification method that used Radial Basis Function<br />
(RBF) classification.<br />
Figure 4-3 shows the RBF classifier. A vector to be classified is passed to a set of basis<br />
functions, each returning one scale value ϕ j j=1, 2,…,l. The concrete choice of the basis function is<br />
not critical; the common implementations prefer the radial basic function (the Gaussian “bell”<br />
functions). Hence the name: RBF network, with :<br />
j<br />
j<br />
ϕ (x) = e<br />
j<br />
2<br />
D<br />
−<br />
2<br />
2σ j<br />
ϕ 1<br />
+<br />
F 1 (x)<br />
X<br />
ϕ 2<br />
+<br />
F 2 (x)<br />
Choose<br />
Largest<br />
F j (x)<br />
j<br />
ϕ l<br />
λ lk<br />
+<br />
F k (x)<br />
Figure 4-3: RBF architecture for classification<br />
68