Thesis (PDF) - Signal & Image Processing Lab
Thesis (PDF) - Signal & Image Processing Lab
Thesis (PDF) - Signal & Image Processing Lab
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
66 CHAPTER 5. TREE SEMILATTICES<br />
Figure 5.5: An example of a projection of the mapping function. Mapping functions<br />
M1(x) and M2(x) of point x are projected to a subtree t1 ∧ t2.<br />
Proposition 2. The tree representation infimum is given by:<br />
T1 ∧ T2 = (t1 ∧ t2, Pt1∧t2(M1) �t1∧t2 Pt1∧t2(M2)), (5.3)<br />
where t1∧t2 = Cr(t1∩t2), Cr(·) stands for the connected graph containing r, and Pt ′(·)<br />
means the projection onto t ′ , i.e., the closest vertex in the path linking the operand to<br />
r, which belongs to V (t ′ ).<br />
In words, the tree representation infimum is built in the following way: The<br />
intersection of the trees is calculated, and the sub-tree containing r is extracted. This<br />
is the infimum of the trees. For each point in E, the mapping function is obtained by<br />
calculating the infimum vertex of the projections of the original mapping functions<br />
onto the infimum tree.<br />
Proof. The infimum must have the following properties:<br />
T = T1 ∧ T2 is infimum iff :<br />
1. T � T1 and T � T2.<br />
2. If there exists T3 such that T � T3, T3 � T1 and T3 � T2, then T3 = T .<br />
The graph part of the tree transform is created by a regular set intersection and<br />
Cr(·) operator, thus t ⊆ t1 and t ⊆ t2 . The mapping function of T is calculated from<br />
infimum of vertices in the tree t1 ∧ t2 , so M � M1 and M � M2.