Pages: [1]   Go Down

Author Topic: Check this out ...  (Read 2748 times)

Jeremy Payne

  • Guest
Check this out ...
« on: May 18, 2009, 11:29:09 pm »

http://acquine.alipr.com/

Interesting idea. The site has some kind of algorithm to "score" the "Aesthetic Quality" of a photograph.

This one scored a 92/100 ... the highest of all 50 I entered ...



... but this one scored 7.4/100 ... go figure.  I like this one better ...

Logged

Eric Myrvaagnes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 22814
  • http://myrvaagnes.com
    • http://myrvaagnes.com
Check this out ...
« Reply #1 on: May 19, 2009, 12:03:04 am »

Maybe subtracting their score from 100 will give more "accurate" results.  
Logged
-Eric Myrvaagnes (visit my website: http://myrvaagnes.com)

bill t.

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3011
    • http://www.unit16.net
Check this out ...
« Reply #2 on: May 19, 2009, 03:30:37 am »

"The Mona Lisa"...26.9
"Moonrise Over Hernandez"...40.3
Thomas Kinkade "Candlelight Cottage"...84.2
Logged

Jeremy Payne

  • Guest
Check this out ...
« Reply #3 on: May 19, 2009, 06:40:28 am »

Quote from: bill t.
"The Mona Lisa"...26.9
"Moonrise Over Hernandez"...40.3
Thomas Kinkade "Candlelight Cottage"...84.2
A friend submitted a pure color field and it scored a 36! They clearly have some work left ...  

On a serious note ... I've always wanted an algorithm to rank images by sharpness/focus and wondered whether that was doable.
Logged

Jonathan Wienke

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5829
    • http://visual-vacations.com/
Check this out ...
« Reply #4 on: May 20, 2009, 10:17:45 am »

Quote from: Jeremy Payne
On a serious note ... I've always wanted an algorithm to rank images by sharpness/focus and wondered whether that was doable.

That's fairly easy, actually. All that is necessary is to evaluate the amount of high-spatial-frequency information present in the image data and calculate a score for the image, then sort accordingly. The only real trick would be separating noise from actual image data, or high-ISO images might score a false positive for focus. You'd probably have to segregate images by ISO (only compare results with other images shot at same ISO setting) for best results.
« Last Edit: May 20, 2009, 10:20:08 am by Jonathan Wienke »
Logged

DarkPenguin

  • Guest
Logged
Pages: [1]   Go Up