Computer vision has today less buzz but much more real and practical application. I’m a little impressed with the state of the art of this field. I remember those days at the Universidad La Salle when some friends and I was trying to get some advantages in image processing with the “new” mmx functions of the Intel Pentium processors and when we try to implement some Douglas-Peucker Approximation with the aid of the voodoo graphics cards and the glide libs. Today you don’t need almost anything about mathematics and you just need a standard computer to do very impressive things.
I think the most important thing of all is that today there are some good companies and startups that are making great business when before it were just an academic interest.
Microsoft for example always had some research on CV, and has some libraries out there, but now is making business with a more small company that has the kinetic technology. This is a good juncture I think and a good opportunity for startups or companies that clearly has a single-technology focused strategy.