Slashdot Mirror

← Back to Stories (view on slashdot.org)

Face Recognition On Mobile Phones

Posted by timothy on from the no-more-anonymity dept.

gpvillamil writes: "This article describes a collaboration between Motorola, Visionics and Wirehound to build in an automatic mug shot recognition capability into mobile phones. Particularly interesting is how the phones will scan all faces in the field of view, and indicate matches by an instant short message."

4 of 131 comments (clear)

  1. Biometrics by govtcheez · · Score: 2, Informative

    That's good - that means I can wear a hat or sunglasses and the police won't be able to catch me. The current state of biometrics is nowhere near good enough for this anywhere, let alone in cell phones.

  2. Re:Who will have access? by CProgrammer98 · · Score: 2, Informative

    Th earticle says that it's for use by law enforcement agencies, not for private phones.
    -

    --
    And the people shall be oppressed, every one by another, and every one by his neighbour Isaiah 3:5
  3. Visionics... by Znork · · Score: 3, Informative

    Is a BS artist company. The products do not work as advertised. The problem isnt in the products tho, it is in the advertising.

    Facial recognition works fine for security checks, where your face gets compared against a small database and the recognition software can be trimmed to rejecting people. The physical settings like lighting can provide the best possible conditions for the software to provide an accurate match, and people can try again if they get rejected.

    It also works fine in finding possible matches for a photo in a database. Now you can trim the software to a high acceptance rate, and get a bunch of likely matches which you can sort manually.

    But it does not work if you need to compare a large database like wanted libraries against a massive number of people because you cannot have it both be certain to trigger on the people it's interested in but not the people it doesnt want. You get a minimum of false positives and negatives which will become most of the triggers when you have a large dataset. A device which will be wrong almost all the time isnt useful.

    Of course, the actual article is rather fuzzy about the use. If it's used for scanning suspects at the site to speed up police work, it will actually be useful. If it's used for scanning everyone the police passes it wont be. And with Visionics being involved in various of the spectacular post 9/11 'facial recognition' projects, I wouldnt be surprised if they attempt to pull another bullshit job.

  4. Salon has an article about this. by Anonymous Coward · · Score: 1, Informative

    Check out the Salon article about this. This is also a good writeup.