I have made several prior blog entries on the amazing resolution of giga pixel cameras. Today I checked the state of the art with a google search. The result is that the event giga photography appears to have advanced as well as the quality of the photography. I would imagine that facial recognition could identify maybe 60 to 80% of everyone in the picture with fairly high reliability. It would not take more than the facial recognition in iPhoto.
What amazes me is the surveillance application. While these are sporting events, any event that is equally well lit with people generally looking in one direction is subject to the same technology.
Blakeway is an excellent example of the application. Blakeway uses mega pixel high res cameras to stitch together pictures. The stitching is excellent.
Giga pixel pictures taken with high res cameras of sporting events at this link.
Note that there is a difference between stitching together pictures taken by a mega pixel camera and a single picture taken with a giga pixel camera of a stadium section when the object is people moving. The same could be done with a mega pixel camera and some degree of telephoto. Giga is mass capture all at one time.
A simple test of facial recognition is to run the algorithm on one picture and then apply the results to some subsequent pictures to find the degree of accuracy. A further test to apply faces from one game to find the same faces attending a different game.
The application of this technology to mass facial recognition is enormous.
There is more than meets the eye here in terms of a business model and application harvesting and data mining this giga pixel information.
Consumer gigapixel cameras are coming.
No comments:
Post a Comment