Intuitive basis for linear transformations (2D/3D matrices)

92 Views Asked by At

Can every possible linear transformation, at least in 2 or 3 dimensions, be expressed as a simple sequence of scaling, rotation, stretch-squeeze, and reflection?

1

There are 1 best solutions below

0
On

Every linear transformation corresponds to a matrix and we know that we can decompose a matrix into its SVD decomposition.

$$A = U\Sigma V^T$$

The orthogonal matrices corresponds to rotation and reflection.

The diagonal matrix corresponds to scaling, stretch-squeeze. However, note that sum singular values can be zero. Hence this corresponds to projection to lower dimensional space.