If you're interested in ELA, I'd encourage you to try it on some control photographs you've taken yourself, and know are unmodified, particularly photographs that have objects with considerably different textures, like textured fabrics.
Trying it on lossless screenshots can also be amusing.
ELA does work when someone manipulates an existing JPEG. If you take a low quality JPEG, paste in part of a high quality JPEG and save in high quality you can easily tell which area has been modified.
Performing ELA on the output JPEGs when the editing has been performed on the high resolution raw images in Photoshop does not work. The down-sampling process effectively normalizes the noise levels.
The ELA analysis in this article is pure nonsense, it is simply highlighting the areas with high frequency components such as the dress texture, fine hairs and features on the face.
To do a proper study like this you need to compare against a sample of similar images which you know have not been manipulated.
In this case, isn't there a sort of "end justifies the means" applicable to this analysis, coming from the fact that the author now has access to the (closer to) original photo?
The apparently shoddy job (handbag, tiles, etc) suggests to me that whitening dark skin tones, enlarging busts and thinning arms and legs is so banal it's not even worth paying a skilled artist for.
To me, this example serves as a stark reminder that women's bodies are still expected to conform to absurd norms in the media. FFS, can we get past this already!?
Both men and women of all kinds are used, not as norms, but as fantasy in art and advertising. And it works because people are attracted to this fantasy and the very few who can attain something close to it.
Natural attraction is not something we can just "get past" unless you have a way to change human nature.
People respond to incentives. They must've figured out at some point that using attractive models for advertising sells better to normal, average people. Not all models you have on hand will be attractive enough for the magazine, so they can supplement with Photoshop. Even if you regulate away Photoshop usage in the marketing/advertising industry, they'd just have to start only hiring super attractive models right? Maybe we could force companies to use average looking models? I don't know, but there's a lot of ways you could approach it. The simplest but most difficult solution is to change people. I know the first reaction is to think that "Wow, it's all those pervert men's fault for these absurd norms and expectations of women!" but then I remember these are advertisements targeted to women. What to do?
And they still find it advantageous to edit their images... touch-up skin tones to make them more "generically white", bigger boobs (when most of the models have ample bosoms already), remove "flab" from elbows. And then remove nipples, because despite selling lingerie and pushing a sexy image, a cloth-covered nipple is a bridge too far.
It's a sad testament to Western perception of beauty and body norms.
C19 has been estimated to cause a 5% dip in global production. Given how controversial this small drop has been, suggesting that people be comfortable with their natural selves is likely to receive some amount of push-back. (it'd affect not only "beauty" and "fashion", but probably also some fraction of "automotive")
> C19 has been estimated to cause a 5% dip in global production. Given how controversial this small drop has been, suggesting that people be comfortable with their natural selves is likely to receive some amount of push-back. (it'd affect not only "beauty" and "fashion", but probably also some fraction of "automotive")
I don't mean to be sarcastic, but I genuinely can't get it. What does "be comfortable with their natural selves" have to do with C19, or a dip in global production?
If people weren't yearning for unrealistic images (note: this is neutral as to whether the irrealism ultimately comes from demand pull* or production push) the impact on GDP from foregone purchasing would probably be a substantial fraction of the currently observed C19 impact.
Beauty: 2% UK consumer expenditure (1,8% UK GDP)
Fashion: 7% UK consumer expenditure
Transportation: <14% UK consumer expenditure
(as noted only a fraction of the automobile part of this might be a more generally male equivalent of fashion)
(a decade ago, there seemed to be two distinct types of car brochure: those that show all the happy people inside the car, and those that show the car from the outside, with reflections avoiding any hint of the occupants)
* as noted above, even the soviets, not otherwise known for producing a wide range of consumer goods, had fashion mags.
100% disagree. Male models have similar unhealthy standards, there's just no social expectation that they shouldn't.
A comment elsewhere lists common conditions for male models. They have to pre-pump before pictures so muscles are more visible, dehydrate to increase vascularity, etc. That comment doesn't mention the editing in post. Hairlines and teeth are corrected in post, hands are increased in size, forearms are warped to match the hands, waists are squashed to give an unrealistic hip/chest ratio.. pretty much an exact analog to what we do for female models.
I think it's a bit myopic to say this is an "all lives matter" scenario, we have social pressure to reduce the pressures put on women in the media. No such push-back exists for men (yet) despite body dysmorphic disorder affecting men and women equally.
Great comment. It's also kind of weird that people get hung up about the photoshopping, when literally nothing about how those photos are taken is "real".
You're taking someone who is already the top 1% of physical beauty (who may have fasted / dehydrated themselves), giving them professionally styled hair and makeup, flattering lighting, making the clothes fit more perfectly than they ever could normally (e.g. pinned at the back out of sight)...
...then having an expert photographer take 1000s of photos and only choosing the top 1% of those photos. Before you even think about post processing!
It's definitely kinda weird but it's bizarre that anyone sees it as an expectation. It's a fantasy.
We are in many ways simple creatures. We see images and respond. And now we throw our greatest minds at work to add machine learning into the mix, to extract that consumer potential from us all to the fullest.
Yes models are expected to conform to an absurd level of physical perfection, just like athletes are expected to have an absurd level of physical fitness, and engineers an absurd level of technical skill. It's their job.
The parent comment is not talking about the model’s actual appearance, but rather what is done to them artificially above that. The correct analogy in athletes would be doping.
Up until about now, actors have been more like "athletes without doping". (Of course with make-up and great attention to lighting, though still you had to look good for an entire scene.) But now with technology which can de-age and deepfake like features, that boundary is pushed even further.
> To me, this example serves as a stark reminder that women's bodies are still expected to conform to absurd norms in the media.
When male actors and models are seen in films or photographs they are shaved or at least heavily trimmed, they have just worked out to cause a "pump" in their muscles, they are dehydrated to make them appear more lean and vascular and they are perfectly lit from above to show off the definition. How's that for an absurd norm?
Not only that but women have the power to choose where to shop and what to look at. They are not poor creatures who need your help. Who do you think buys from Victoria's Secret and reads the women's magazines? Women do. Let them choose.
It's worth re-mentioning that the model in this article had her skin tone whitened significantly as part of the photoshopping, as well as nipple removal/bust enhancement/handbag "removal" etc.
What exactly is he performing the PCA on? I assume the image is concatenated into a 1D vector, so what makes up the matrix that's used to generate the eigenvectors? RGB values?
I may be getting this wrong, but doesn't that slide show the PCA being done on multiple images? Each image is 321 x 261, which concats to a 83781 vector. Then they're repeating that for 32 images producing a 83781 x 32 matrix, which they then apply the PCA on to get 32 principle components.
When you have only one image, what's making up the columns of the matrix (assuming rows = 1D image vector)?
> I may be getting this wrong, but doesn't that slide show the PCA being done on multiple images? Each image is 321 x 261, which concats to a 83781 vector. Then they're repeating that for 32 images producing a 83781 x 32 matrix, which they then apply the PCA on to get 32 principle components.
Yes, you are right (except that it's 'principal'), and I didn't read my link carefully. I am a mathematician whose specialty is not a million miles away from these kind of signal-reconstruction techniques, but far enough away that I can't give any better answer from personal knowledge. Sorry!
Not on topic, but the changes to the pics bring out the horror masks. Check the black skull leering out of the gloom in 2nd pic, the joker's red/green skull-face of the 4th, the black cheshire cat grin of the last.
This is interesting when it comes to being able to spot a Photoshop. But this seems to be about more than that. It seems to be about casting shame on to Victoria's Secret for using Photoshop. Why is that? They are selling a dress and want to show it in the best light. Is the author claiming false advertising? The model is being compensated (probably handsomely) for the use of her image. She should consider herself lucky that Photoshop exists because she's clearly not exactly what they want, but close enough.
If you're interested in ELA, I'd encourage you to try it on some control photographs you've taken yourself, and know are unmodified, particularly photographs that have objects with considerably different textures, like textured fabrics.
Trying it on lossless screenshots can also be amusing.