[Jun8]

Maybe I missed in the article but it seems they're not using 3D modeling. Why not, esp. considering the prevalence of cheap 3D sensors? According to Google there are many people working in this area (as well as shoe fitting).

The goal should be to totally smash current custom fit prices, e.g. to ~$50 level, e.g. using a self-serve fitting station at Macy's. Many such stations can be monitored by one person, similar to how self-serve check out stations are monitored in grocery stores.

The above is for mass market. OP's manual approach can still be used as a high-end service for those who can afford it, e.g. pret-a-porter vs. tailor.

[kaitai]

I agree with most of your comment, but the difference between bra fitting and shoe fitting is that you don't want to change the shape of your foot. Most women do want to change the shape of their breasts, even if not consciously: for instance, if your breast point outward & spread a lot, the fit of clothing is changed; if you have droopy breasts, you generally want to hold them up, again in the interest of clothing fit if nothing else; if you're involved in sports, you may want to create certain shapes to optimize your participation.

3D fitting would be really interesting, but not straightforward. It's hard to do the nonlinear transformation that would be necessary, partly because density of breast tissue varies a lot and so not all transformation functions would actually be possible. The "bloat" factor outlined in the article would also have an effect -- hormones, salt consumption, water consumption all affect that. 3D modeling might be most useful for placement of underwires, etc.

[pureliquidhw]

I'm guessing because there are more dimensions than simply shape of breast that affect the fit of a bra.

I wonder if a machine could be made that inflates/deflates a proxy of a bra until the woman (wearer? Am I assuming gender?) is comfortable. Take the readings of a comfortable fit and construct a bra that provides that support. Basically, you don't care about the shape of the breast, you care about the shape of that individual's comfortable breast and the pressure needed to get it into that state.

[ynniv]

Optical measurements are good for rigid objects, but humans are squishy. Accurate measurement for clothes requires some interaction and judgement. We could eventually have robots do it, but cameras alone aren't enough.

[downandout]

She said in the article that they abandoned the 3D modeling idea after the results turned out to be less than ideal. I also cannot really understand how they couldn’t have worked such a thing out though, even given all of the other factors she mentions in the article that they must account for. It would seem that all of the factors could fairly easily be thrown into the algorithm.

[pbhjpbhj]

AIUI breast size and shape alter, even for fixed body weight. Full 3D modelling might be an over-optimisation?