Complex phenotypes like face shapes aren't trivially determined by a single gene . Instead, the face is formed through the interactions by many gene products, and the differences in genotypes between individuals contribute in highly non-linear ways. For many years people tried to apply fairly simple reasoning (first just mendelian, then later SNPs, then copy number, then SNVs) on small numbers of genes but you can't explain all the hereditable genotype/phenotype variation without the whole genome (and possibly the epigenome).
For faces they generally noticed a bunch of changes in a subset of genes that are involved with embryonic and later development. And specific patterns of changes across multiple genes.
It's generally believed now that we could sequence millions of people and take normalized photos of their faces (see eigenfaces), sampling roughly from global population clusters in a balanced way, and build a joint genotype/phenotype embedding that would allow us to probe those complex, multivariate, non-linear rules. You would be able to do prediction and generation (given this genotype, show the most probable phenotype/face, or given this face, give a plausible genotype). Unfortunately this would be a highly politicized process makign it challenging to complete.
Your quest makes you concerned with discerning from averages. Outside your world, people don't have to be concerned with averages. Instead, they can see multiple correlated features. Perhaps a jawline, a brow, and a hairline.
That said, though, you've typed something so unless you're sitting for the Turing test soon I'd guess you do resemble me more than a box turtle does. Knowing that features are influenced by genes, the idea that similar features imply similar genes is a very simple hypothesis to make. I'm a little puzzled why it's so interesting to test whether genetics and epigenetics influence appearance in 2022.
To me, it's near meaningless unless they start with a very strict definition of resemblance. Is it based on some measurement or someone's subjective interpretation of who looks like whom? I'm not paying $4.25 a week to read TFA and I've been told I "look like" plenty of people I don't. There are two people I know of in the world who resembled me very closely at some point in my life. One was fairly local and I disbelieved he really looked so much like me until my own aunt approached him in public and he said he'd heard the same about me. The other was a guy I knew from a BBS (actually a talker, if anyone here knows what that is) who decided he had a crush on me and sent me a photo of himself. I printed it out, held it up beside my face, and my mother asked when I had that photo taken. Freaky resemblances do exist. Should I be surprised that the people who looked like me are closer genetically and possibly in traceable genealogy than another human of another gender, another nationality, another ethnicity, another culture, and another society?