What would actually be useful is something that does the same for food, although I have no idea how that'd work (perhaps a probe of some kind). It's trivial to calculate nutritional information for beverages as a human, but for those of us who track such things, I have yet to see a really good system for solid foods that doesn't rely on human guessing about portion sizes and matching to existing known entities from a database of foods.
Computer vision researchers are starting to work on the food calorie counting problem. See this video demo from SRI International: https://www.youtube.com/watch?v=rH2c_F6YB-A
Why is it trivial to calculate nutritional inoformation for beverages? In the United States, on average, 21% of our calories are consumed from beverages [1]. I know some people that just drink water, so that means there are others who are consuming probably greater than 1/4 of their calories from beverages. There is clearly a problem, in the USA, of people consuming too many calories from beverages, and it is not trivial. I think this is a good step, but there are a few ways it may be missing the mark (like why not a set of mugs that I can change the user settings of based on who in my family or friends grabbed the cup out of the cupboard.)
Yeah, thought that too. Some approximation on calories and macronutrients would be really great.
This would definitely serve the huge market of folks who track their daily intake:
- People who (have to) do it for medical reasons
- People that want to loose weight and need to get a better understanding of what exactly they are eating every day (often times it is not enough or the ratio of macronutrients is out of place)
- Nearly every bodybuilder and fitness-addict
What would actually be useful is something that does the same for food, although I have no idea how that'd work (perhaps a probe of some kind). It's trivial to calculate nutritional information for beverages as a human, but for those of us who track such things, I have yet to see a really good system for solid foods that doesn't rely on human guessing about portion sizes and matching to existing known entities from a database of foods.