Basic Mathematical Tools

for Imaging and Visualization

winter term 2022/23

PD Dr. Tobias Lasser

Computational Imaging and Inverse Problems

Technical University of Munich
Program today

1 Linear Algebra
Linear spaces
Subspaces and linear spans
Coordinate systems and bases

2
Linear spaces
Linear space (over R)
Let V be a nonempty set with the two operations a + b ∈ V and λa ∈ V for every
a, b ∈ V and λ ∈ R.
V is called a linear space if the following rules are fulfilled:
(V1) a + b = b + a for all a, b ∈ V (commutativity)
(V2) (a + b) + c = a + (b + c) for all a, b, c ∈ V (associativity)
(V3) there exists a zero element 0 ∈ V such that a + 0 = a for all a ∈ V
(V4) for every a ∈ V there exists an inverse element −a ∈ V such that a + (−a) = 0
(V5) 1a = a for all a ∈ V (1 ∈ R real number)
(V6) λ(µa) = (λµ)a for all λ, µ ∈ R, a ∈ V
(V7) λ(a + b) = λa + λb for all λ ∈ R, a, b ∈ V
(V8) (λ + µ)a = λa + µa for all λ, µ ∈ R, a ∈ V

3
Program today

1 Linear Algebra
Linear spaces
Subspaces and linear spans
Coordinate systems and bases

4
Parts of linear spaces

Subspace of a linear space

Let V be a linear space. A nonempty set U ⊂ V is called subspace of V if
• x, y ∈ U =⇒ x − y ∈ U,
• x ∈ U, λ ∈ R =⇒ λx ∈ U.

Remarks:
• a subspace U of V is a linear space itself.
• short notation: U ≤ V .

5
Subspace properties
Let V be a linear space.
• Let Ui , i ∈ I , be subspaces of V , then i∈I Ui is a subspace of V .
T

• Let M ⊂ V , then there exists a smallest subspace ⟨M⟩ of V that contains M.

⟨M⟩ is called the (linear) span of M in V .
We have \
⟨M⟩ = U
M⊂U≤V

(intersection of all subspaces U containing M) and we have

( )
X
⟨M⟩ = λx x : λx ∈ R, almost all λx = 0 ,
x∈M

(“almost all” means all except finitely many λx are 0), i.e. ⟨M⟩ is the set of finite
linear combinations of elements of M.

6
Generating linear spaces

Generating set
Let V be a linear space. A set of vectors (bi )i∈I in V is called generating set of V if


(bi )i∈I = V .

7
Program today

1 Linear Algebra
Linear spaces
Subspaces and linear spans
Coordinate systems and bases

8
 Tracking
Augmented Reality
Tracking system

PatientCT

Probe

Patient

11

10
Independence

Linear independence
Let V be a linear space. A finite set of vectors a1 , . . . , an ∈ V is called linearly
independent if

λ1 a1 + . . . + λn an = 0 =⇒ λ1 = . . . = λn = 0.

A set of vectors (ai )i∈I in V is called linearly independent, if all finite subsets of (ai )i∈I
are linearly independent.

11
Coordinate systems generalized

Basis
Let V be a linear space. A set of vectors (bi )i∈I in V is called basis of V , if it is a
generating set of V that is linearly independent.

12
Basis properties
• Every linear space V has a basis.
• Let b1 , . . . , bn be a basis of the linear space V . Then every a ∈ V can be written
as a unique linear combination
n
X
a= λi bi with λi ∈ R.
i=1

• Let V be a linear space. Then b1 , . . . , bn ∈ V basis

⇐⇒ b1 , . . . , bn ∈ V minimal generating set of V
⇐⇒ b1 , . . . , bn ∈ V maximal linearly independent set in V .
• Let V be a linear space with basis b1 , . . . , bn . Then every basis has exactly n
elements.
This invariant is called dimension of V , dim V := n.

13
Images

14