Reader Comments

Subscribe: RSS

View by: Time | Thread

• 

  dan, 15 Dec 2005 @ 10:38am

  
  

  Precentages not degree, but degrees of belief

  The percentages are not parts. That is, 80% happy, 20% sad does not imply that she is part happy, part sad. These percentanges denote what degree of belief we should assign to the propositions
  
  . Mona lisa is happy
  and
  - Mona lisa is sad
  Therefore, it is 80% certain that she is smiling, but there is some probability that she is Perplexed. etc.
  

  [ link to this | view in chronology ]

  • 

    azuravian, 15 Dec 2005 @ 12:54pm

    
    

    Re: Precentages not degree, but degrees of belief

    If that is the case, then couldn't the total percentages add up to more than 100%, since they don't represent opposites. For example, it would be reasonable that someone, based on a study of their face, etc. the degree of belief for them being happy could be 80%, while also having a 35% degree of being confused.
    
    Is it then, just a coincedence that in this case, the degrees of belief all add up to 100%.
    
    Or, I could just be wrong.

    [ link to this | view in chronology ]

    • 

      stephen pray, 15 Dec 2005 @ 1:05pm

      
      

      Re: Precentages not degree, but degrees of belief

      Not that anyone cares what anyone else thinks naturally, but personally it seems to me the emperor has no clothes.. since i was a kid i haven't seen any smile on her face. everyone talked about the enigmatic smile.. but to me it appeared to be benign neutrality. smile? where? but then i have never seen a man in the moon face either...

      [ link to this | view in chronology ]

• 

  denacopoliez, 15 Dec 2005 @ 10:48am

  
  

  79% interested

  I always feared this day when we wouldn't be able to enjoy things like art anymore because the passion has been replaced by science. This proves to me that People are too busy doing absolute nothing! But still this make me 58% worried and 22% hungry.

  [ link to this | view in chronology ]

  • 

    Stephen Tillman, 15 Dec 2005 @ 11:44am

    
    

    Re: 79% interested

    My interest in this comes in at the source of the need of this program. I read the article that talked about "the enegmatic smile". Ok... what's the mystery. I've never understood what the big deal is with her smile. It's not like the art of the times depcited big, toothy grins.
    
    Any art/history/humanities majors out there know what the great mystery is/was all about?

    [ link to this | view in chronology ]

  • 

    PLogo, 15 Dec 2005 @ 12:09pm

    
    

    Re: 79% interested

    I hope you're not really personally offended by this research as your post would suggest. This exercise in recognition served as a nice demonstration of the possibilities of AI (artificial intelligence). Obviously we don't NEED a computer to tell us that Mona Lisa is probably happy, but the fact that a computer could figure that out for itself is a brilliant achievment. The group that did this is probably not the first or only to advance facial recognition, but they had an amusing way of bringing their work into a brief spotlight. I don't understand why to insist on dragging it through the mud without even trying to fully understand its implications. Go ahead and feign annoyance like a baby once again. Oh, and stop pretending that science is weaking the institution of Art.

    [ link to this | view in chronology ]

    • 

      denacopoliez, 16 Dec 2005 @ 6:08am

      
      

      Re: 79% interested

      PLogo: I think that you might need to switch to Decaf, Trust me everything I said was with a huge dose of sarcasm, and I am glad the a computer was able to figure out that she's happy, but i dont need one to tell you you sir need to loosen up. I didn't read any bashing here it was just humor.

      [ link to this | view in chronology ]

• 

  Joe Snuffy, 15 Dec 2005 @ 11:12am

  
  

  Where can I get one?

  Do they make a program that can tell what your girlfriend / wife is feling? Now that would be of some use.

  [ link to this | view in chronology ]

  • 

    me, 15 Dec 2005 @ 11:25am

    
    

    Re: Where can I get one?

    I can tell you how she's feeling. ;)
    
    sorry...couldn't resist.

    [ link to this | view in chronology ]

  • 

    I Own Posh!! Dangit !!!, 17 Dec 2005 @ 4:46am

    
    

    ThEE OuT BUrsT!!! I GO OuT- POETIC

    Whats next... is the goverment using this technology to try to convict suspects.
    And... should we really..put all are... trust and hope in this science.. after all... science is always saying its right... then proving itself..that its wrong ...its true. I remember when people were excited about the lie detector test... and turns... to be faulty one. I have a friend who has passed a lie detector test on lying alone...what does that tell you... are bodies... are..too complex! Science alone cant...absolute the human emotion with "Yeeah..I am scientist...I know that I can detect your human emotion with my machine to the last digital percent.. ah doeey..ada da da."..Really..lets not play-ourselfs.......when science claims its right...with anything associated with complex matter and numbers..it always come back with a different number. You all been great..Thank You Thank You Very Much. Have a nice day.
    
    

    [ link to this | view in chronology ]

• 

  Plogo, 15 Dec 2005 @ 12:07pm

  
  

  somewhat overzealous critique...

  I hope you're not really personally offended by this research as your post would suggest. This exercise in recognition served as a nice demonstration of the possibilities of AI (artificial intelligence). Obviously we don't NEED a computer to tell us that Mona Lisa is probably happy, but the fact that a computer could figure that out for itself is a brilliant achievment. The group that did this is probably not the first or only to advance facial recognition, but they had an amusing way of bringing their work into a brief spotlight. I don't understand why to insist on dragging it through the mud without even trying to fully understand its implications. Go ahead and feign annoyance like a baby once again.

  [ link to this | view in chronology ]

  • 

    John, 15 Dec 2005 @ 12:20pm

    
    

    Re: somewhat overzealous critique...

    Man. Someone needs their humor sensors rejiggered again. Dude, it was funny. The only one acting like a baby is you.

    [ link to this | view in chronology ]

• 

  nunya bidness, 15 Dec 2005 @ 1:02pm

  
  

  100% amused

  :)

  [ link to this | view in chronology ]

• 

  Novicane, 15 Dec 2005 @ 1:14pm

  
  

  Emotion recognition is very important!

  I think the context of this could be shifted more to a technical aspect. Software with the ability to recognize facial expressions, and associate human emotion in percentages could be extremely useful in developing more user friendly software. Software that could judge the response of an individual, or the emotion set in a query could provide better results or more helpful insightful assistance with whatever job it is doing.
  
  If someone is angry and software can recognize this, particular colors, tones and keywords could be used to influence the person’s mood and ease their anger. Or if a person is depressed or unhappy, the software could adapt and try to find a way to help the individual feel better. All of the attempts to alter the mood of the individual could be tracked and measured also by real time processing of those persons emotional responses. Psychology teaches us that colors, sounds, wording, and vocabulary has a major impact on our moods.
  
  Testing this kind of software on media that we know to be a display of a particular emotion, as smug as it may be, could really help test and ensure the software is able to read said emotions to a greater degree of accuracy.
  

  [ link to this | view in chronology ]

  • 

    Mike (profile), 15 Dec 2005 @ 5:26pm

    
    

    Re: Emotion recognition is very important!

    I think the context of this could be shifted more to a technical aspect. Software with the ability to recognize facial expressions, and associate human emotion in percentages could be extremely useful in developing more user friendly software. Software that could judge the response of an individual, or the emotion set in a query could provide better results or more helpful insightful assistance with whatever job it is doing.
    
    Indeed. But, testing it against a painting seems fairly useless. There's simply no way to determine how accurate it is.

    [ link to this | view in chronology ]

• 

  Sv, 15 Dec 2005 @ 6:09pm

  
  

  Next...

  breaking news: the smart expressions analyser software was applied to a number of object in the last few weeks to determine that: Japanese automobiles are happier than US automobiles.
  
  Tommorow: are trees happy or sad?

  [ link to this | view in chronology ]

  • 

    Stephen Tillman, 16 Dec 2005 @ 5:53am

    
    

    Re: Next...

    You know... Japanese auto's do look happier. Hmm... wonder if that's a cultural thing.
    
    Trees though... gotta be kinda miffed that they have to stand there for thier -whole- life. Kinda take the chipper outta me.

    [ link to this | view in chronology ]