5 ms·

I feel that whenever Lena is mentioned, we should also give a shoutout to the researchers who were fed up with her and decided to use Fabio instead. https://ar

I feel that whenever Lena is mentioned, we should also give a shoutout to the researchers who were fed up with her and decided to use Fabio instead. https://arxiv.org/abs/1202.6429 https://arxiv.org/abs/1202.6429 What do you guys (yes, guys) think about making Fabio the standard image instead?

qubex 8y ago
Well on her (the researcher), her (Lena), and him (Fabio). Well on all of them.
- brokenmachine 8y ago
  You mean "good on them"?
johndough 8y ago
The Fabio image is not very useful wrt image compression: - grayscale (lenna image is rgb) - very odd 271x257 resolution (lenna image is square 512x512 which is a power of two) - just hair as fine detail features (lenna image has several: hat, feather boa and hair) https://4.bp.blogspot.com/-jUHXZ0jW4Lc/T1t3W0ex2BI/AAAAAAAAFeY/7TY0cTty1og/s400/Fabio.jpg https://4.bp.blogspot.com/-jUHXZ0jW4Lc/T1t3W0ex2BI/AAAAAAAAF... http://www.betapleatedchic.com/uploads/3/0/4/1/30413754/4907429_orig.gif http://www.betapleatedchic.com/uploads/3/0/4/1/30413754/4907... An optimal image would have the following properties: - different textures at different - scale - brightness - frequency - easily recognizable features - color - power of two resolution - no cultural issues - no licensing issues (not a scan of a magazine, globally recognized license, i.e. not "public domain") The problem with faces is that on the one hand, humans are able to tell easily if a feature was destroyed by an image compression algorithm, but on the other hand it's personally identifiable information and therefore comes with additional legal restrictions in many jurisdictions.
craigsmansion 8y ago
> we should also give a shoutout No, we shouldn't. Them using a different image in their research is perfectly acceptable--when one spends hours looking at something rendering, one shouldn't have to look at something one objects to--but referring to the research as a social statement diminishes the scientific value by framing it as a publicity stunt. Also, apart from using research time/money for licensing: "lena" has become shorthand for "face-female-with-various-patterns" and is a historical accident. "Fabio" will always be Fabio, a commercial model; a less than optimal choice for a scientific paper.
taneq 8y ago
I'd have no issue with it, but then I have no issue with the Lena image. Of course, anyone who does have an issue with Lena would have equal issue with a male equivalent, if they're not a hypocrite.
- cedex12 8y ago
  > if they're not a hypocrite. I don't agree at all with your statement here. There is a whole historical/social difference at play here; taking it into account doesn't make you a hypocrite. (I hate the "hypocrite" judgement. So easy to cast and rarely true)
  - taneq 8y ago
    Well I hate the way that "it's OK to do something to a person of one race/gender/subgroup that it's not OK to do to a person of a different race/gender/subgroup, for some historical reason" has become the new justification for all sorts of prejudice, so I guess we're even.
    - yesenadam 8y ago
      Please don't do this on HN.[0] >Of course, anyone who does have an issue with Lena would have equal issue with a male equivalent, if they're not a hypocrite. Part of cedex12's point, I take it, is that there is no 'male equivalent'. Your 'of course' following by accusation of hypocrisy makes it seem you were looking for a fight, not a discussion, and mainly wanted to dump out the standard 'talking point' which followed. "Of course" is just bluff, begging the question, code for "I have no reason to give but I'm right, obviously". No, you're not "even". (I do however disagree with cedex12's claim that accusations of human hypocrisy are rarely true.) [0] I dislike like the sound of me saying that. I tried removing it, but I disliked the implied acceptance of the way you were speaking on here even less, so I put it back.
ThrowawayR2 8y ago
> What do you guys (yes, guys) think about making Fabio the standard image instead? No objection here. Heck, make both the Fabio and Lena images part of the test image set so that everybody's happy (or at least equally unhappy).

qubex 8y ago

Well on her (the researcher), her (Lena), and him (Fabio). Well on all of them.

brokenmachine 8y ago
You mean "good on them"?

johndough 8y ago

The Fabio image is not very useful wrt image compression: - grayscale (lenna image is rgb) - very odd 271x257 resolution (lenna image is square 512x512 which is a power of two) - just hair as fine detail features (lenna image has several: hat, feather boa and hair) https://4.bp.blogspot.com/-jUHXZ0jW4Lc/T1t3W0ex2BI/AAAAAAAAFeY/7TY0cTty1og/s400/Fabio.jpg https://4.bp.blogspot.com/-jUHXZ0jW4Lc/T1t3W0ex2BI/AAAAAAAAF... http://www.betapleatedchic.com/uploads/3/0/4/1/30413754/4907429_orig.gif http://www.betapleatedchic.com/uploads/3/0/4/1/30413754/4907... An optimal image would have the following properties: - different textures at different - scale - brightness - frequency - easily recognizable features - color - power of two resolution - no cultural issues - no licensing issues (not a scan of a magazine, globally recognized license, i.e. not "public domain") The problem with faces is that on the one hand, humans are able to tell easily if a feature was destroyed by an image compression algorithm, but on the other hand it's personally identifiable information and therefore comes with additional legal restrictions in many jurisdictions.

craigsmansion 8y ago

> we should also give a shoutout No, we shouldn't. Them using a different image in their research is perfectly acceptable--when one spends hours looking at something rendering, one shouldn't have to look at something one objects to--but referring to the research as a social statement diminishes the scientific value by framing it as a publicity stunt. Also, apart from using research time/money for licensing: "lena" has become shorthand for "face-female-with-various-patterns" and is a historical accident. "Fabio" will always be Fabio, a commercial model; a less than optimal choice for a scientific paper.

taneq 8y ago

I'd have no issue with it, but then I have no issue with the Lena image. Of course, anyone who does have an issue with Lena would have equal issue with a male equivalent, if they're not a hypocrite.

cedex12 8y ago
> if they're not a hypocrite. I don't agree at all with your statement here. There is a whole historical/social difference at play here; taking it into account doesn't make you a hypocrite. (I hate the "hypocrite" judgement. So easy to cast and rarely true)
- taneq 8y ago
  Well I hate the way that "it's OK to do something to a person of one race/gender/subgroup that it's not OK to do to a person of a different race/gender/subgroup, for some historical reason" has become the new justification for all sorts of prejudice, so I guess we're even.
  - yesenadam 8y ago
    Please don't do this on HN.[0] >Of course, anyone who does have an issue with Lena would have equal issue with a male equivalent, if they're not a hypocrite. Part of cedex12's point, I take it, is that there is no 'male equivalent'. Your 'of course' following by accusation of hypocrisy makes it seem you were looking for a fight, not a discussion, and mainly wanted to dump out the standard 'talking point' which followed. "Of course" is just bluff, begging the question, code for "I have no reason to give but I'm right, obviously". No, you're not "even". (I do however disagree with cedex12's claim that accusations of human hypocrisy are rarely true.) [0] I dislike like the sound of me saying that. I tried removing it, but I disliked the implied acceptance of the way you were speaking on here even less, so I put it back.

ThrowawayR2 8y ago

> What do you guys (yes, guys) think about making Fabio the standard image instead? No objection here. Heck, make both the Fabio and Lena images part of the test image set so that everybody's happy (or at least equally unhappy).