Details
In this code, the mask for

is {{0,0,1},{1,1,1}}. Below, the shapes are given instead of the arrays as a visual aid for the reader:
This function is very specific to the overlap tilings seen in the resource function FindMinimalTilings.
A tile mask is a subset of positions within a rectangular array, such as the
a values in
{{a,a,a},{_,a,_}}. This particular mask is also known as the Tetris T shape,

.
The results of
FindMinimalTilings on mask

has identical results to rotations and reflections of

.
The results of running
FindMinimalTilings on mask

has identical results to mask

(skew equivalency).
Similarly,

and

are equivalent masks for overlap tiling systems.
Evaluating
ResourceFunction["CanonicalTilingMask"] on any rotation or reflection of masks

,

or

returns

.

,

and

are canonical.

is the canonical form of

.

is the canonical form of

.

is the canonical form of

.
Currently, many hours of computer time are needed to run
FindMinimalTilings on masks

,

and

. These are all equivalent to

. The point of this function is to find these equivalencies before an expensive run.
For overlap tiling systems, these results only meaningfully apply to 2D shapes. The domino array {{1,1}} can be considered a 1D mask.