The Future of Beauty Tech: AI, Innovation & Smart Solutions with Wayne Liu

Summary

Coffee N° 5 with Lara Schmoisman
Episode: The Future of Beauty Tech: AI, Innovation & Smart Solutions with Wayne Liu
Release Date: February 25, 2025

---

Introduction

In this insightful episode of Coffee N° 5, host Lara Schmoisman engages in a comprehensive conversation with Wayne Liu, CEO of Perfect Corp., a pioneering company at the intersection of beauty and technology. As a trailblazer in beauty tech, Liu shares his journey, the innovative technologies his company employs, and the transformative impact of AI on the beauty and wellness industries.

The Genesis of Perfect Corp.

Lara Schmoisman (B) welcomes Wayne Liu, expressing her admiration for his work and introducing him as the CEO of Perfect Corp., a leader in beauty technology. She emphasizes the revolutionary nature of his company's approach in merging beauty with technology. Wayne Liu (A) delves into the origins of Perfect Corp., highlighting the merger of beauty and technology over the past decade. He recounts how, faced with the stagnation of PC growth, his team pivoted to leverage their expertise in color, video processing, and photo processing to create innovative solutions for the beauty industry.

Wayne Liu [01:17]: "We merged beauty and technology, starting with the idea to make the most frequently used mobile feature—the camera—more beautiful and fun."

Innovation in Virtual Try-On Technology

Liu explains the foundational idea of leveraging mobile cameras as mirrors to enhance the user experience. The company’s initial focus was on solving common challenges faced by consumers, such as uncertainty in applying makeup and the friction retailers face in consumers trying new products.

Wayne Liu [04:28]: "We are solving the problem from both the consumer and the brand side by reducing the friction of trying new things."

Business Model: B2C and B2B

Perfect Corp. operates on a dual business model, catering to both consumers (B2C) and businesses (B2B). Initially, they launched a suite of apps under the YouCam family, experiencing rapid organic growth with millions of downloads. While the apps remain free, premium features are available through subscription fees. On the B2B side, Perfect Corp. licenses its virtual try-on and skin analysis technologies to over 700 brands, ranging from luxury names like Chanel to mass retailers like Walmart.

Wayne Liu [05:00]: "Our customers are both the brands who license our technology and the consumers who use our apps to enhance their beauty routines."

Expanding into Fashion and Accessories

Liu discusses the expansion of Perfect Corp.'s technology into fashion and accessories, addressing the unique challenges this posed. Transitioning from facial applications to full-body accessories required advanced technologies like Physically Based Rendering (PBR) and AI-driven 3D modeling.

Wayne Liu [07:03]: "We use PBR to simulate object movement with lighting, enabling accurate virtual try-ons for accessories like watches and shoes."

To further enhance their capabilities, Perfect Corp. recently acquired Wana, a leader in virtual shoe try-on technology, allowing them to offer a comprehensive virtual try-on experience from head to toe.

Wayne Liu [09:14]: "With the acquisition of Wana, we can now provide virtual try-ons for shoes, complementing our existing facial and accessory technologies."

Overcoming Technological Challenges

Ensuring accuracy and realism in virtual try-ons was a significant challenge. Perfect Corp. developed color blending technologies to accurately represent makeup colors across diverse skin tones and finishes. This meticulous approach ensured that virtual try-ons were not mere stickers but realistic representations that brands could trust.

Wayne Liu [11:18]: "Our color blending technology ensures that makeup colors reflect accurately on different skin tones and finishes, making our virtual try-ons truly real."

Embracing AI: Present and Future

AI has been integral to Perfect Corp.'s innovations from the outset. Liu shares his perspective on the continual evolution of AI, emphasizing its non-reversible integration into technology advancements. He underscores the importance of understanding and leveraging AI responsibly to solve real-world problems.

Wayne Liu [15:26]: "The question is not whether to use AI, but how to use it effectively to solve problems."

Medical Accuracy and Partnerships

Perfect Corp. extends its AI capabilities into the medical realm through skin analysis tools. Collaborating with universities and medical professionals, they ensure that their AI-driven diagnostics are accurate and unbiased. Liu highlights a study conducted with Dr. Steve Feldman, a board-certified dermatologist, which validated the accuracy and consistency of their skin analysis tool compared to traditional methods.

Wayne Liu [17:02]: "Our collaboration with Dr. Feldman demonstrated that our AI is not only accurate but sometimes less biased than traditional methods."

Addressing Concerns: Self-Diagnosis and Ethical Use

Liu addresses potential concerns about self-diagnosis using AI, emphasizing the importance of organizational support and responsible usage. While Perfect Corp.'s tools empower users with information, they are designed to complement, not replace, professional medical advice.

Wayne Liu [24:00]: "Our technology empowers users but is not intended to replace professional medical diagnoses or prescriptions."

Future Possibilities

Looking ahead, Liu envisions further integration of AI in areas like drug discovery and advanced diagnostics, positioning Perfect Corp. to continue leading innovations that bridge beauty, wellness, and technology.

Wayne Liu [25:18]: "This is just the beginning. With proper harnessing of AI, we can generate even more benefits and transformative solutions."

Conclusion

Lara Schmoisman concludes the episode by expressing her fascination with the advancements and future possibilities presented by Wayne Liu and Perfect Corp. The conversation underscores the pivotal role of AI in revolutionizing the beauty and wellness industries, offering listeners a glimpse into the future of beauty tech.

Lara Schmoisman [25:31]: "I agree with you 100%. Thank you again for having coffee with us today."

---

Discover More:
For additional insights and resources discussed in this episode, visit larashmoisman.com or check the episode notes below. Don’t forget to subscribe to Coffee N° 5 for more actionable business strategies and inspiring success stories.

Transcript

[A]

Foreign.

[B]

This is Coffee Number Five. I'm your host, Lara Schmoisman. Hi everyone. Welcome back to Coffee Number Five. And I'm sure you're wondering what we're going to be talking about today. Because I always have such an amazing guest and interesting topics and unique topics and that it's basically what interests me. And I'm so happy and humbled that interests you as well. And you became a listener of Coffee Number Five. I wanted to introduce you today to Wayne Liu. How are you, Wayne? Welcome.

[A]

I'm doing very well. Thank you for having me.

[B]

You are incredible. I met you a while ago and I was so impressed. I mean, I hear you talk here and there and I wanted to have you here because you have to be here. You are the CEO of the Perfect Corp. The Perfect Corp. Is beautytech. And you're the first one in this space. And I really want to understand how did you come up with this revolutionary idea? But not only that, then from going from the idea to implementation when nobody else was doing it.

[A]

Yes. So first of all, I am very happy to talk about the journey of Beauty Tech because we are actually getting the intersection of a beauty and technology. So 10 years ago we merged these two together and then right now it's a merger, these two perfectly together. That's just like our company. So actually everything start from I would say just it's the idea because you know, me and my. Which is the CEO Alice. Alice who always. She's in Taiwan. So. So that's why, you know, we. We work together actually in our private company, previous company, we work together as a company called Cyberlink. So Cyberlink is basically software company doing the software for the PC. But you know, at that time like a PC is sort of like a stop growing, which, you know, so that's why we are. It's about 10 years ago, so we have some ideas. We have a tons of technology, especially on color and the video processing, photo processing. How can we apply all this technology? Make it fun. Okay, so then we start doing the brainstorming, all kinds of things. First thing is people are using a mobile phone. They are using mobile phone. But the most common application they are using the mobile phone is a camera. Okay. And then actually pretty significant of this female, they are using camera as a mirror. Right? So that's why, that's how every.

[B]

I mean, you can say that I did this before as soon as we came into the meeting. I really like my lipstick.

[A]

Okay. So that's why innovation usually start from this kind of brainstorming. Okay. We want to get into this the most frequent used features, which is a camera mirror. When people see this as a mirror, we want them to see them more like a beautiful and fun. So that's why apply our technology like especially in the color photo processing and then start doing this virtual color. Okay. So you can actually we are solving problem because you know, female one of the most challenging thing is putting makeup and then take out putting makeup. So that's why we are using technology to help to solve this problem. In the meantime, it's fun.

[B]

So yeah, it's so challenging to know that you don't know how a makeup is going to look on you.

[A]

Yes, absolutely.

[B]

Resolve a problem and saying. And because it's not only about putting the makeup, it's learning how to put the makeup. It's not something that it comes with us. Applying makeup is difficult. So even if you're gonna try something that doesn't look good on you, you're gonna feel horrible about it.

[A]

Yeah, absolutely. And also from brand and retailer perspective, it's a. Because the more you try, the more you buy. However the thing is if you create a friction of trying like every time you need to put on your faces and wipe it out and for many different reasons, including like a hygiene reason like you know, you don't wanna get into other people's this kind of tester. So that's why we are sort of solving the problem from both the consumer and the brand side because we reduce the friction of trying new thing. So you can now put into a green lipstick to see how it looks. But usually you won't do that. But right now with a virtual technology you can definitely do that. That's how we started. Yeah. So 10 years ago.

[B]

I have a question for you. So then who is your consumer? Who is your client? It's anyone can go out there and use because one thing is for you. Okay. I created this technology but then was to understand who is my client here. So who is your client?

[A]

So we are not only innovated on the technology, but we also is very innovative on the business model. So our company, we actually has two sides of business. One is a B2C and another one is B2B. So for the actually we start with the B2C. We create an app. Right now it's a multiple app. It's just not one, but it's a YouCam family, you know, YouCam cam like a camera. So we have the app when we started the app is just organic growth. So we see the download like 100,000, 1 million, 5 million, 10 million. So people just love it. They start doing that. So that consider our B2C customer. However, you know, previously we don't really charge them. It's a free app. But now actually with a lot of AI features there's a premium, it's still free, but if you don't want to access some premium content, they can definitely do that by paying the subscription fee. So that's one form of business which is a B2C. Another one is B2B. So that's pretty much I would say the similar technology like a virtual try on or the skin analyzer we put in our app, we can actually license to the brand we currently work with the more than 700 brand, just the world renewal brand like SE Lauter Co. You know, like all these LVMH brand, even the luxury brand like a Chanel all the way to the the mass retailer like a Walmart. So they license our technology. You may not know this ours but if you go to Walmart's website they see. Okay, so if you want to try elf lipsticks, okay. So you can just click the website and try it. And then if you go to Chanel's website, you want to try something. As I said, not only just the makeup, but we actually get into Chanel's fashion like a watch or something. You can try it virtually is powered by us. So technically our customer is both a brand and the consumer. We are using different model business model to get into these two.

[B]

And now you mentioned something very interesting that you are also working to fashion and accessories.

[A]

Yes, yes.

[B]

So how difficult was to move from the technology that it goes only from your face to do accessories that it gets in different part of the body and with different body types.

[A]

That's a great question because there really poses technology challenges. First is in our cases we start with the faces. So everything on the faces. Of course there's some also probably have a pretty significant amount of the accessories on your faces like an earring and eyewear. So that one is fine. However, if we go into some other body part like you try to try a watches or wrist, that's a new technology. So that's why like a traditional virtual on doesn't work. We have a. We have another set of technology called the physical based rendering PBR which we try to simulate the movement of the object with a lighting. So like a light tracing. So not without getting to too technical but basically that's how it works. Okay. So we are using. Yeah and also we are using some Innovate technology and also AI to recreate the 3D models. Okay. So like we call the authoring. That's how we get into that.

[B]

Okay. So if for example, I want to work with you as a brand, you need to recreate each one of my products as a rendering to be able to put it in the technology. Or this is works with a photo.

[A]

Yes. So there are different level of engagement. For example, if you just have a quick solution. So usually we do a 2D to 3D. So basically you just give us photos. We can recreate the 3D model for you, of course, because it's from 2D to 3D, so it's great. However, it's not as, I'll say precisely accurate as a real 3D model. Some of this luxury brand, like a really luxury brand, they really care about all the detail, reflection, material. So in that case, we pretty much recreate, we need to create a 3D model from some of their 3D files. So that's why we start to rendering.

[B]

So I'm thinking like for example, in a case of shoes, how would you do that? Because you need to point your camera in the bottom of your feet. It's not in front of your camera on the face. So how do you work with another body parts?

[A]

Yeah, so for the shoes, it's a, again, that's a different story because of. We technically we don't have a shoe solution before we acquire a company called wana. We acquire Wana like a one month ago. So that one is really the company One Hour is a leader in the, in this virtual trial on fashion, especially on shoes. So of course like our virtual try on technology, it doesn't apply to the shoes. So that's why we acquire one. So that 1R technology sort of like a compensate what we have. So now from face, actually we have hair from hair to toe. Okay. So they have their unique technology to do the, to do the virtual try on for shoes. So that's why. Yeah.

[B]

Okay. This is mindblowing because how do you do it in stores? Probably you can have a big mirror and you can do it and you can see full.

[A]

Just so like a. It's a mobile app, right. So it's just like you have a, you have a mobile phone. You can, you can just place the mobile phone in front of your, your, your food and then you know, the virtual try on show up on top of your.

[B]

I know, but how can you see it? If you put it, do you get a picture of it or how you see it?

[A]

Because if you lower, it's a real time. So it's pretty much like you are using a front facing camera to point to your shoes. And then you are actually from the first person angle you look at the camera, it's just like you are taking a photo of your shoes.

[B]

I see. So I see.

[A]

Yeah. So it's a similar. It's just like you are for example, you just take a photo on top of your own shoes or food. Okay. And then on top of that, we just change it into the whatever shoes you want to try. Yeah.

[B]

Okay, that's great. So what did you find as a more challenging of creating all these applications? What was the thing that you say, oh, are we gonna pull this off? Are we gonna able to do it?

[A]

Yeah. So actually different activation, different product has a different challenge. For example, let's talk about our original like the color. Okay. So makeup, color, the first thing is you have to be true. Okay. True to real. Otherwise if you just like a sticker. Right. So it become a more like a fun tool. It's not really a useful tool for the brand can use it. So that's why first thing is like you know, years ago, we are trying to overcome the color. The color is not just a color because the first thing is the color. If you put the same makeup, the lipstick on different people because your native tone. Yeah. So the final finish actually is different. So that's why we have our pattern technology called color blending. We understand your skin tone, lips and then we blind with the color and also we think about the reflection, different finish. Like sometimes you can get Metallica, you can get matte. So that's why we are playing around the magic of technology to reflect the different color, the texture. So that's why, that's how we start to conquer all this challenge and then become a real. So when the brand see this final finish. Yeah. So they say, wow, that's real. It's not just a sticker. It's real.

[B]

Then you have to work with real life models with different tones in order to get accurate in the.

[A]

Yeah, so that's one of thing which we are really proud of. Yes, we did. We in our cases, actually that's AI. That's machine learning. We can understand through all this machine learning with all this working with a brand and with a customer, we actually can understand 89,000 different skin tone. 89,000. It pretty much together in entire spectrum. Entire spectrum of the color, the skin tone. So that's why let say talk about inclusivity. Right. So can you create 89,000 different color for your customer. Probably you don't. But the thing is, yes, we can identify.

[B]

So that's why did you at the beginning did you think that they're going to be so many colors?

[A]

No, it's just, you know, gradually it's impossible to develop so many. I mean the skin tone to understand. So that's why gradually we work with customer. Some of the customer, they don't really, they don't have this wide range. They only cover a certain part of the spectrum. But some of the customer like Mac, they really have a lot of color. So that's why we need to, we need to sort of like working with them. That's how you know, working with the customer try to understand their need, try to solve their problem actually in the meantime enhance our technology. So in the past 10 years that's how we getting to like every time we have a challenge we find a solution and we get it ourselves into the another level.

[B]

Yeah. So let me ask you a question because these, I mean you guys were pioneers working with AI technology. So what do you think is next? Because AI in the last two years they've been booming. Do you think that now slow down or are we going to be able to see new things that you are like really impressive.

[A]

I. Yeah. So actually we've been using AI from day one because you have to understand like a face. Right. The location. Right location. So that's why we are, we have a extensive understanding of AI from always I said it's a classic AI supervised learning. That's why we learn all the skin, you know, skin tone, skin condition and then there's a machine learning so supervised to. Now it's a non supervised which usually call it generative AI from my perspective or from the customer. We've been working with the industry. I think this process is not reversible. So now it's not like asking will the AI fading or will AI stop or post. I don't think it will. So the question how are you challenging.

[B]

AI to take it to the next level.

[A]

So that's why I say the question is not should I use AI? The question is how should I use.

[B]

AI exactly anymore people need to use AI AI because of our life.

[A]

Couple of things for using AI is I think first thing is you have to have some kind of understanding of AI. We know it's just going too fast. But the thing is knowledge is really the key. If you have no idea about AI even the simple AI, you don't even know what generative AI is different from machine learning, the classic AI. I think that would be very difficult in a creative field because you don't know everything. You don't know, you have uncertainty. That's why you get a little bit feel. So once you get a little bit comfortable about AI, get some knowledge. I think that's really the thing. And then you need to have a motivation to turn this AI knowledge into the skill. What do I mean by that? Motivation pretty much from like us, from a technology provider's perspective is really to solve problem. We are not using AI to create more problem, we are using AI to solve the problem. So just in our cases, like you know, AI skin analysis. Analysis. Right. So we are using AI to understand your skin in five seconds. In five seconds, which we empower the customer. It can be the medical service provider, it can be a brand. So we help to understand your customer and your customer understand their skin. So that's how we using AI to solve problem. So that's why you need to have a, you have knowledge, you have to have a motivation. Yeah. And then. Yeah.

[B]

How accurate is this medically?

[A]

Okay, that, that's question whether that's one of the challenges. It was a challenge, but now it's not. So because we've been asked many, many, many, many times on this, how accurate is your skin analysis? So okay, then we decide, let's do this because everything is a scientific base. So we work with the university, you know, very fair third party. So we start establish this research project with the university. So for example, actually if you're interesting, we send you some of this paper. We work with Dr. Steve Feldman who is a professor and also a board certified dermatologist of Wake Forest University. So we've been working with him like four years ago. So he conduct a very thorough research on our product. So basically that's a simple way to put it, is using AI to diagnose one customer, one client and using professional tool which they've been using for years in their industry, diagnose the same customer, same condition and then have a panel of a doctor, the physician to do a general assessment. Now we have a three, we have an AI, which is us. They have the devices they've been using, they have a doctor. So he compare this three of course using a statistic and all kind of thing and come down with a result, very fair result to see are you accurate, are you consistent compared to this? Yes, the answer is yes. So that's why he has a, you know, he published paper on the medical journal. Okay, so that's a Peer reviewed general. So this means. Okay, then you're accurate. Then next is a will. And then he start using our tool to analyze the different condition. And then one thing which people ask about beside the accurate next is. Is your AI bias. Okay. So then he conduct another research he just published in the one of this medical journal early last year. Okay. So early last year March. So they use a Fitzpatrick to understand six different type of skin. And then same thing using ours using machine and using doctor and compare. Actually our score is higher than the current machine. And then sometimes because he check a different category. Some of the category actually AI is the less biased than the doctor. So that's why he's very excited with all this data. He published a paper.

[B]

What he's saying is basically that now we can use technology as a partner for doctors.

[A]

Of course. Yeah. So it's not. It's people has been using it now. Okay. So we have in our website we have lots of a case study actually. Doctor, dermatologist, plastic surgeon, they are using our tool so for. Yeah for kind of like some of them. One of our early partner, Dr. Park, Eunice park, she's been using for almost three years in her practition.

[B]

And how do they feel how a physician will use these tools that will help them to be more accurate. They help them to optimize time. Why a physician will use this technology.

[A]

Yeah, so different like individual using different way. But in general they're using couple of things. First is help them to do an early diagnostic. Okay. Because when the patient come in, you know right now this we short of people in the. In the medical like service healthcare people know that. So that's why instead of has a probably a front desk or nurses to ask some question, we are using AI like a 5 second. We have a thorough understanding of a skin. Okay. So like 14 conditions we have a different. We are using numbers, numerical numbers. And then the doctor can. When Dr. It's about time to see the patient is pretty much safe their time to do the consultation. Because just like you know check like.

[B]

X ray when you go like to the eye doctor and you got.

[A]

Yes, yes, yeah. Yes, of course. That's the one thing. That's one thing. And then also the second thing is it gives a more confidence for the patient because when patients step into the doctor's office usually they get. They are vulnerable because they don't know what happened. Okay. So by understanding like visually overlapping their faces, their own faces not model their own faces with the skin with a different. We're using color Coded like a different color coded to highlight your skin condition and then numerically measure your skin condition. You pretty much understand. Okay, now I have a thorough understanding like my skin. And then that's easier for Dr. To start to have a conversation. And also because some of them using this stimulation function again that's a. Like a. Through tons of the, the learning the data. So we do simulation. Okay, so what if, so what if you do this treatment, your condition will become like that. So this kind of a conversation is really help sort of empower the medical service provider and also empower the customer. The client. Okay. And then from brand, okay, so from the brand perspective, some of these brand, they are using this one to do product recommendation because we have an algorithm to sort of like imagine the condition with the product. Okay. So that's why you can do a knowledge transfer. You don't. You may not have a very experienced staff but because the knowledge already been installed in the, in the, in the, in the app. So that's why you can do a.

[B]

Very good recommendation in about off the counter solutions. Yeah, so we are. So this would not be for a derm. They cannot use it for prescription?

[A]

No, no, we are not getting to that. We are not getting to the so called like FDA approval pilot thing. Okay. So we are not really doing any like a treatment but for the brand they can recommend product, their product but not really like from the doctor they are recommend the treatment because the treatment is a pretty sensitive word on the, you know, on the medical term. Not yet. So potentially maybe not now we are pretty much, you know, focused what we've been doing now not getting to like a FDA type of.

[B]

Do you think that in the future this is something that you see possible using this technology?

[A]

Yes, definitely. And also like a drug discovery and also like a diagnosis like a disease. It's some of not our company but some other companies doing that with the AI. So especially on the drug discovery. That's a very popular application of AI.

[B]

Yeah, but how risky is this for people to self diagnose with issues? You mean is this a concern that people can use this application and then self diagnose with problems?

[A]

Well, it really depends on like previously I talked about knowledge and motivation and the last part is really the organization support. What I mean by that is of course we have a technology for you to do self diagnostic. But the thing is how are you going to use it? Okay, so what are you going to allow your customer to do this and then start to get a prescription drug? No, of course it's not the cases but if you want to use have your customer or your client or patient like doing the following. Okay, so one of the things actually when we talk to the doctor, one of the problem they have is a customer disappear their patients after one treatment. They disappear because it's probably come to this location or something, it's too difficult or two challenges, they don't have a time too. So that's why if you can do a remote diagnostic, try to follow. Okay, so monitor following the patient's progress. I guess that's a good way to do it. However it's like okay, using this self diagnostic and then start to figure out your own problem and then get the prescription. I guess it's not within the current framework.

[B]

Wayne, thank you so much. This was incredible. A lot of knowledge. Thank you. I'm fascinated with all this technology and where it's going to take us and the possibilities that we see in the future.

[A]

Yeah, yeah, of course. And then I believe it's just the beginning. And then, you know, with a good, I would say the control harness of AI AI will generate more benefit for us. Yes, absolutely.

[B]

I agree with you 100%. Well, thank you again for having coffee with us today.

[A]

You're welcome. Thank you to you guys.

[B]

I see you next week with more coffee.

[A]

Number five.

[B]

Find everything you need at larashmoisman.com or in the episode notes right below. Don't forget to subscribe. Was so good to have you here today. See you next time. Catch you on the flip side.

[A]

Ciao, ciao.