I like the scheme where you rectify the transformer to provide power to a 'proper' gate driver IC while also using the transformer signal as the input to the IC. This circuit is a 'clever' variation on that general idea.
On the other hand when I've studied the problem I haven't come up with a reason to use a discrete transformer versus a modern isolated DC-DC plus an isolated driver IC.
Mass produced royer type DC-DC's are just about as cheap as discrete gate drive transformers as far as I can tell for my applications.
Vicor used (maybe still does) a simple custom IC to get
indefinite hang time from a pulse xfmr driven gate. A fat
forward Schottky for charge, and once charged, a second
"tickler" pulse applied through the same xfmr triggers a
2-transistor SCR to discharge the gate (too skinny to
give a spurious turnon). Seems maybe a bit noise sensitive
but they never indicated a problem with it.
I had a good look at the guts as I was supposed to do a
redesign of it, before the program ran out of money. They
also told me their "pulse transformer" was really just a pair
of minimum spaced traces on the module substrate (like
maybe more of a "transmission line balun" before that was
a commonly understood thing).