In theory you can. I was involved in a cryogenics project some years ago which demonstrated the measurement of mass flow using two in line capacitance sensors. Each sensor measured the % of liquid versus vapor within the volume of the capacitors (short plates installed in the flow stream). The dips in capacitance represented high vapor areas passing through the plates. The time series signals from the two sensors were then time correlated to give a time shift and knowing the distance between the two sensors, a velocity. The mass flow is the density times the area times the velocity. The density came from the average signal of capacitance which is proportional to the mass density (times the dielectric permittivity). The problems that we had related to the collapse of vapor pockets and the changing character of the two phase flow. I would think that this would be easier for milk and air (basically water and air) but testing would need to be performed and it probably works best for a range of flow rates with intermediate amounts of bubbles.

Two-phase flow is extraordinarily difficult to measure, as the two phases will be moving at different speeds globally and locally. That is, there will be two 3D velocity profiles (one for each phase) and these will vary over time, even at a constant total flow rate. The only way to know for sure is to measure the two phases individually before they mix or combine them and measure them after they become one, in which case you will need to be able to distinguish between the two using something like a tracer (chemical, radioactive, etc.). Much work has been done to characterize the flow patterns and make rough estimates of the slip, void fraction, etc. Over the past 38 years I have on a number of occasions been called upon to measure the flow of wet steam and have tried many different techniques, all rather disappointing. Don't promise anyone you can measure two-phase flow to closer than +/-10%. In fact, it's difficult to measure single phase flow to within +/-5%, although I have achieved better than +/-1% on several occasions, using dye dilution. If someone tries to sell you a "flow meter" and claims it's accurate to +/-5%, they're lying or they don't know what they're talking about. Acoustic flow meters are nowhere near as accurate as the manufacturers claim--perhaps under ideal conditions, but not in the field. People who deal with dye dilution gloss over most of the uncertainty. Hand waving and ignoring uncertainty or averaging a whole lot of inaccurate numbers doesn't make the result any more accurate. You're lucky if a Pitot tube is good to +/-2% under conditions and fluid similar to those used in the calibration. Of course, a Pitot won't work for two-phase flow. If the flow was always similar and you could calibrate it, a drag plate or annubar might give somewhat consistent results. But how could you calibrate such a device in the first place? Also don't be too hopeful about Doppler devices because you won't be able to get enough information fast enough and the assumptions you will have to make can't be verified or even tested, so you will just be making up numbers. I'm sorry to be so discouraging, but measuring two-phase flow almost qualifies as a quest for the holy grail.