I see two options:
1. A pattern of dual pixels across the sensor (in addition to the ones for focusing?) which can take a deep depth of field and simulate a shallow one. While Canon has the technology for this (and this is definitely the cooler of the options), it feels too techie for them to use in entry-level gear (where most whiz-bang features tend to be software).
2. Focus stacking with lots of fps. This seems more likely to me, especially if they market this as a macro option (since macro shots still can get kinda shallow DoF on a compact). Also, it's mostly a matter of software (provided there's processing power for it), which is more consistent with consumer products.