Study Web

Linear Transformations

Linear Transformations

Definition of linear transformations, standard matrix representation, 2D transformations (rotation, reflection, scaling, shear), composition, invertible transformations.

Linear Transformations

Definition: A function T: ℝⁿ→ℝᵐ is a linear transformation if:
1. T(u+v) = T(u) + T(v)  (additivity)
2. T(cu) = cT(u)  (homogeneity)
Equivalently: T(cu+dv) = cT(u) + dT(v)
Matrix representation: Every linear transformation T: ℝⁿ→ℝᵐ can be written as T(x) = Ax for some m×n matrix A (the standard matrix of T).
Column j of A = T(eⱼ) where eⱼ is the j-th standard basis vector.

Standard 2D Transformations

TransformationMatrix AEffect
Rotation by θ (CCW)[cos θ, −sin θ; sin θ, cos θ]Rotates vectors anticlockwise by θ
Reflection about x-axis[1,0; 0,−1]Flips y-component sign
Reflection about y-axis[−1,0; 0,1]Flips x-component sign
Scaling[s,0; 0,s]Scales by factor s
Horizontal shear[1,k; 0,1]Shifts x by k×y
Projection onto x-axis[1,0; 0,0]Sets y=0
Rotation by 90° CCW: A = [[0,−1],[1,0]]
T([1,0]ᵀ) = [0,1]ᵀ ✓ (x-axis maps to y-axis)
T([0,1]ᵀ) = [−1,0]ᵀ ✓ (y-axis maps to negative x-axis)

Composition of Transformations

T₂ ∘ T₁ corresponds to matrix product A₂A₁
(apply T₁ first, then T₂)
Order matters: A₂A₁ ≠ A₁A₂ in general

Invertible Transformations

T is invertible iff det(A) ≠ 0. Then T⁻¹ corresponds to A⁻¹.
T⁻¹ undoes T: T⁻¹(T(x)) = x

Kernel and Image

  • Kernel (null space): ker(T) = {x | T(x) = 0} — inputs that map to zero
  • Image (range): im(T) = {T(x) | x ∈ ℝⁿ} — all possible outputs
  • Rank-nullity: dim(ker T) + dim(im T) = n