Transcription
Criteria that I used for mate selection for both sexes, and for the majority of animals, it's up to debate as to whether love at first sight truly exists. But I think we all know it's a bit more obvious when it comes down to attraction at first sight: you're either attracted to them or you're not. With that in mind, individuals who are physically attractive tend to have more options when choosing a mate and have more opportunities to pursue a good quality mate. This is one of the underlying tenants of the beauty and cosmetics industry, who seek to identify and market the cues and predictors of facial attractiveness to make those tips and tricks accessible to a larger population.
Researchers have identified a number of cues in facial health and attractiveness, including averageness, symmetry, sexual dimorphism, or things that we've talked about in the past. But there is also an additional concept that we haven't discussed, which is the value of a skin condition and the importance of recognizing facial health and attractiveness through the quality of your skin. Facial skin texture and, more importantly, color play an important role in observers' judgments of apparent health and have been linked to aspects of physiological health, including fitness, immunity, and fertility. Existing research has confirmed this phenomena for Caucasian populations, with smooth and bright skin indicating fertility, fitness, and health, which has led to the popularity of foundation and blush in the makeup industry, which aim to mimic exactly that: a homogenous skin smoothness through foundation and increased blood flow through blush.
Now, we've talked about the makeup aspect of this topic in more detail in our video on why the Egyptians wore eyeliner and the video on blushing can make you more attractive. However, in this video, we're going to put more emphasis on the contribution of skin texture to the perceived health of faces, and instead of a Caucasian population, we're going to look at the other races, starting off with an Asian sample by Tan et al. 2017. This particular paper, which was funded by the schools of psychology from Southeast Asia and Australia, explored the contribution of skin texture and color to the apparent health of Malaysian Chinese faces. As background research, the researchers recognized why skin smoothness and color have such a strong correlation to raters' judgments of health. It turns out that skin color and smoothness directly correlate to the skin blood perfusion and oxygenation, which predicts physical fitness and cardiac health. This smoothness and rich color in the face also directly relates to the concentration of melanin in the body, which is associated with a healthy immune and reproductive system.
In particular, skin with a sunburned look were not perceived as healthy in Stephen et al.'s study, which looked at Caucasians, and Hannah Tower 2018, which looked at mainland Chinese populations. Unnaturally high melanin, beyond what you'd expect for that person's phenotype due to excess sun exposure, filters out much of the UV that's needed for vitamin D production, and so this lack of vitamin D production is associated with reduced emotional well-being and increased musculoskeletal pain. Much of those with lighter skin tones, say in the harsh Australian outback, such as miners, fall prey to skin cancer, with Australia having one of the highest rates in the world. As that skin tans and becomes leathery, it's going to develop a layer of melanin quite naturally, but this comes about as a result of dangerous levels of sun exposure and, again, reduced vitamin D production when the weather starts to get cooler again. A similar problem is seen in those with darker skin, such as the American Black population, where vitamin D deficiency is especially high for that group. So it's understandable why, in many of these cross-cultural studies, melanin was seen as a perceived health indicator, even though it isn't necessarily true nor accurate for everyone. In fact, there are many facial cues that we use to infer about a stranger's health. However, as Cozy and Tower defined in 2009, it has to actually have a link, or a direct link, to physiological health to be really considered a metric that we can use for facial attractiveness. While we may assume that someone with sunburned skin is not exactly healthy, the burn does heal quite quickly. However, say the sunburn creates a visible melanoma, which then develops into malignant cancer; well, that's a different story, and that indeed is a very visible indicator because it's on your skin. In that case, that could be a negative for your facial attractiveness and your perceived health.
So, with keeping this analogy in mind, let's go back to the original study on skin smoothness. Tan et al. looked at simple correlations between well-defined factors of skin smoothness and their relationship to judgments of attractiveness by raters. As Asia represents a huge diversity of cultures and ethnicities, the researchers only looked at Malaysian Chinese faces in order to limit the influence of diet and potential activity levels on the results of the study. So 53 students were recruited and had their photo taken; from this, their skin textures were rated on three factors. An independent sample of raters made a judgment on the perceived health and the perceived attractiveness of every face and then made graphs to represent relationships between the amount of a particular skin factor that was displayed and how this affected their facial attractiveness. If you look towards factor one, it can be seen on this graph that larger values of this certain factor directly relate to a higher perceived rating of skin health, as well as this, the same factor has a very similar relationship to overall facial health, showing how influential the skin can be in an overall judgment of the face. But just what is factor one? If we look at the bottom set of cheeks here, the top left cheek has the highest values of factor one, and the bottom right cheek has the lowest value of factor one. As all of this is measured on what's known as a computer-based Gabor wavelet function, which all you really need to know is just another way of measuring and quantifying the skin homogeneity in the pixels of these taken photos. A very dumbed-down way of understanding what factor one represents is just a lack of contrast in the overall height between the bumps and the rest of the skin, or in other words, it's just the most smooth skin tone.
That being said, skin is one of the most malleable facial features, having the continuous ability to change color and smoothness on an almost daily basis, while also being extremely prone to acne and inflammation. With that in mind, our skin is perhaps the most accessible element of our face that we can readily and cheaply improve upon without expensive surgery. This is incredibly important when we recognize how influential our skin is in these conditional judgments of attractiveness, as we're going to go on to just a second. But before you get into that, it may be worth for you to check out our earlier videos on skin that look at the effects of sun damage in terms of speeding up the aging process, the negatives of acne and inflammation on ratings of attractiveness, and a number of other terms and terminologies that are going to make more sense for this video if you watch those first. While you're doing that, I think you should also hit the like button, or else the bad skin fairy is going to come and get you.
Looking back at our study, there is a significantly large and negative correlation between perceived skin health and factor two of skin texture and a correspondingly significant large negative correlation between this factor and overall facial health. Essentially, the more of factor two that a participant displays, the less attractive that they are rated. Factor two, again, is a similar wavelet function, and it basically represents the amount and the size of the imperfections or blemishes that your skin would have. Once again, quite a seemingly apparent conclusion to draw: that the larger the multitude and size of your imperfections, the lower the skin and face health will be rated. And this is partly because, as we looked at in the previous video on acne, humans are very, very disease-averse. If there's one thing we're scared of, it's diseases, and acne is a bacterial infection, and so we stay away from it. This particular conclusion on having many and large blemishes on your face is quite cross-cultural; I mean, it's a biological thing. Over the three videos that Koovs has made on skin, we've seen very strong consensus in the research that having rough skin is a very big no-no, and a lot of this has to do with disease and pathology and how we've evolved as a species, being very hyper-vigilant to detect skin abnormalities.
The third and final skin texture-related factor that we're going to talk about in this study has a similar but less extreme pattern to factor two. This is a significant correlation between higher values of this factor and lower values of overall perceived skin and facial health; in other words, it's something you don't want to have. If we look towards the diagram for factor three, with larger values represented in the top left corner and lower values in the bottom right, we can recognize that low factor three values represent skin homogeneity, and this is a function of skin color and texture. On the flip side, high amounts of factor three can relate to a lack of internal consistency of the skin itself. When we look at the general factors of facial attractiveness and of averageness of symmetry and dimorphism, it makes sense that consistent and average skin, in the sense that it is smooth, as it's a combination of many different skin patches of different people, then obviously makes sense with the existing literature that's been proposed ever since the 1950s on evolutionary biology and humans preferring disease-free skin. However, there is a silver lining that we can talk about. In each of these studies, there's a continuous outlier that we can see in all of these graphs. Scientific research always seeks to find patterns in the general sample so that they can make generalizations about the larger population, but just like that, scientific research will also never be able to fully recognize the individuality of each participant. These outliers could have had extremely strong and desirable facial features with great proportions, symmetry, averageness, all the other tenants, and could have been lacking in the skin department, and so their rankings were still higher than somebody of lesser attractiveness in terms of facial features but not as good as if somebody just like that but with perfect skin as well. What this indicates to me is that skin health is indeed quite important, but it's not the be-all and end-all of facial attractiveness. Being really strong in another department, such as having model-like symmetry or very strong facial features, can make up for not having ideal skin. Also, a lot of people have more access to makeup now, so foundation is worn by primarily still women, but men also have access to it, which can make up for any shortcomings.
With that being said, what do we consider the opposite? Someone with a recessed jaw, very malformed dental facial features, generally poor facial growth, often due to lifestyle choices such as mouth breathing; then, no matter how good your skin is, it's not going to make people still be attracted to you, and there's no amount of makeup that can cover up a missing jaw, a recessed mid-face, very forward protruded eyes. You may think that I'm exaggerating, but we looked at this exact question on your pragmatism where participants pre- and post-surgery were rated on their attractiveness, and unsurprisingly, once these people with misaligned jaws got their jaws corrected, their attractiveness went up by a full three or four points, which is huge. And so if your only cosmetic issue is not having clear skin, then consider yourself quite lucky because it could have been a lot worse. By now, you probably think that skin homogeneity and color are the most important things, but there's actually some other factors to consider as well. Kurti et al. 2012 and Stephen and Tower 2011 considered skin color on apparent health for Black African populations; Tan et al. 2017 looked at Asian populations, while Finger et al. 2011 looked at Caucasian ones, and what did they find? Well, unsurprisingly, the exact same thing, because these skin research tests show strong cross-cultural findings, much more than you'd see for research in other topics like the jaw, the nose—especially not the nose—the forehead, hair; all of those things can more or less differ in attractiveness or how important they are by cultures, but having good skin is very important because it's ingrained into us right from birth.
So we've seen from the findings of Tan et al.'s study and Finger et al.'s and Stephen et al.'s that this very strong cross-cultural agreement—this is something that affects virtually everyone watching this video—but it's for that exact reason that skin is a topic that I'm personally quite passionate about because, for many, it provides the greatest cost-to-benefit in improving your looks, where non-surgical approaches such as creams and even fillers can drastically undo decades of photo damage and poor lifestyle choices, something that no surgical approach can quite achieve. For this reason, we've been working very closely with dermatologists, both in our team and external, to work on a skin care classifier using machine learning models and image segmentation that takes an image of a face, such as a phone selfie; there's a ton of pre-processing to get the skin regions that we need in the accurate resolution, and then has a dermatologist identify everything they can from that patch to train a convolutional neural network. So basically, the dermatologist knowledge is being transferred into a computer-based application. Many of you have outlined your concerns in the poll that we put up earlier, which is great; I'm glad to see so much diverse criticism because criticism is good, especially in the development stage. One of the most important things that I kept saying was the lack of diversity, which is what everyone has a concern about because there are so many skin care applications and machine learning programs that just don't cater to dark skin or Middle Eastern, Latino, so on and so forth. But that's definitely been our top priority in making a very sophisticated clinical classifier and a very high accuracy, but using data sets of Black, White, Latino, Middle Eastern, East, South, and North Asian faces and then combining their features together using a generative adversarial network to get every possible face in between those geographical races. So you know, say your face is somewhere between Mongolian and Russian, then if we combine the two faces, we'll get some kind of a mix between those two. That's the approach that we've taken, and this is a very new approach; I've only seen it in the research literature since 2020, and for us, it's a race against the clock to put out research on that and show that it can actually be used as a viable way of increasing data sets and creating every type of ethnic face possible. GAN classifiers aren't actually new; Ole has one, Ponzi has one, but their versions are very rudimentary and frankly useless because if it's not accurate, there's no point using it. I won't go too much into detail into how exactly we're building ours because we're using a completely different approach to be much more accurate and have much more features, but just for now, just know that it's in the works, and we're hearing all of your concerns, and we're working to make something that's very inclusive, very accurate, and very useful for dermatologists and then secondary for retail users just like you guys. That I'll end the video. So if you're interested in knowing more about this project that we're working on or any of our machine learning and AI projects, you can visit the website and try our skin care version one prototype, which was talked about by MIT, BBC, and a number of different outlets. If you want to visit the website, get Acused Aesthetics report and get your skin assessed by a team of dermatologists and doctors, you can also do that, or you can listen to our beauty podcast where I bring in guests and we talk about different aesthetics issues on Patreon.