Previous | Next --- Slide 60 of 82
Back to Lecture Thumbnails
Samleo

Why are lines and more perplexingly, parallelism, preserved? For lines, is it because the transformation is line-ar?

mpotoole

For lines---yes, it is because all of these equations are linear. Why? Well there's a simple proof for this.

Suppose you're given two points $x_1$ and $x_2$. The line passing through $x_1$ and $x_2$ is given by $x(\alpha) = \alpha x_1 + (1-\alpha)x_2$ for some value $\alpha$.

Now let's apply some affine transformation $A$ to the points on this line. This results in $A x(\alpha) = A ((\alpha x_1) + (1-\alpha)x_2) = \alpha (Ax_1) + (1-\alpha)(Ax_2)$. In other words, we get a line that's passing through two points $Ax_1$ and $Ax_2$.

As for parallelism, that may be a little trickier to prove, but there's a simple justification here. If an affine transformation is a composition of shearing, translation, scaling, and rotation operations, and all of these operations preserve parallelism, then an affine transformation must also preserve parallelism.