Google speeds up image recognition algorithm by a factor of about 20.000

Google Has Another Machine Vision Breakthrough?

The result of the change to the basic algorithm is a speed up of around 20,000 times, which is astounding. The method was tested on 100,000 object detectors based on a deformable-part model which required over a million filters to be applied to multiple resolution scalings of the target image which were computed in less than 20 second using nothing but a single multi-core machine with 20GB of RAM.

Couple this with the Google Glass or something similar and a few powerfull server, and you can search for the "bad guy" in a lot of places very fast.

NSA will be pleased.

reply to post by tyfon


From what I understand, this has a lot to do with being able to identify object classes faster, along with putting more details into the object. This in turn will force the people doing this into somewhat detailing, or fine tuning, the classes themselves thereby making the whole easier to recognize by the system. Easier recognition leads to quicker recognition.

Is that about right?

www.tutorialspoint.com...

Yup, I posted a thread about the actual technical aspects of this sometime before they decided to update their capabilities.

Suffice to say, the technology is being suppressed.

People probably thought I was full of B.S.

Avenged!