 So we're here with Advertima here, and it's doing an interesting shape on my face, and they can analyze and say I'm 32 years old. I'm nearly 32. That's great, so it's very accurate. And you, sharing 26? 22? All right, you're 22 years old. All right, that's what he says. All right, that's cool. So, hello, so who are you? Hi, I'm Iman. I'm one of the co-founders of Advertima, and this is Chris, also co-founder. Exactly. So, this is just one of the things you do, you can detect how all people are? Yeah, I mean, basically what we've tried is to create an artificial intelligence that we're working on, which analyzes the current situation in front of an installation, which can be a screen like we have here. And then, based on this information, it decides what content do I have to show to create a personalized and positive experience for the people in front. So, this is a technology, is it open source or not? How to detect age? Yeah, I mean, it's our own algorithms, yes, exactly. At the moment, we have age and gender, but the next step which we'll have in summer is emotions, tracking, clothes tracking, and as well as eye contact with the screen. Really? No eye tracking, no? Eye tracking at the moment, not yet, but after we will have, if someone actually looked at the screen. You could have an IRB merlin. The idea is you want to put advertising for that person. Yeah, I mean, we're not an advertising company, but it's one of the use cases where we actually implemented it already. So, it's basically a decision-making process, real-time decision-making process. Based on this information, it will decide what's going to be the next clip on this screen, like Man's Toilet, because here we are like four guys now at the moment. It shows content for men. Nice, that's actually amazing, no? Yeah, I mean, it's great because... Has anybody been doing this or not? I mean, the company is trying to do it, but we defined another process, now a new process, which is like tracking, persona analysis, and then the targeting. This is like a unique way how we do it. Now we are talking about screens, but the next output sensors will be speakers, lights, and even spell machines to create a multi-sensory experience. So, are you 59? No. No, that is not far, maybe. Ah, 53. 53, okay, maybe you need to be in the middle of the... It's getting better, look, now it's coming down. All right, so there's a huge potential in using display targeted the right way to the right people. Exactly, I mean... Dynamic. Exactly, in real time. What we have online already for years now is going to come also now into the real world, but we're also working with banks. We also have a case for people flow management, for example in a train station, where you have a lot of people, so the signalatics, the signs, adapt to the amount of people, so also to increase safety within a train station or an airport. But I think it'd be great to just walk around and it kind of like gives you advertising that's a little bit like Google, is trying to optimize advertising, optimizing the display advertising, very important. Exactly, exactly. So, you really just receive this information that is interesting for you, because study showed that more than 80% of the content that we see on a daily basis is not relevant for us. That's why we start to ignore this content and also misinformation that could be interesting for us. Is there any chance that you can, in the future, potentially have a global system or a national system that where people can approve that they would get even personal, even more, and that you would have face recognition for some people who allow you to do it, they would get actually personalized to themselves? We already have the case, actually, because we are working with a Swiss shopping center and there we also implement the app. And in the app, you can upload your profile picture and decide and agree that you will be recognized by your person, by your image. And then you have another open window and a lot of possibilities because you can really target this specific person and retarget it again. And we're also connecting it to the website and the newsletter. So, also, the website is dynamic and the newsletter is dynamic and personalized just for you. So, based on your movements within the shopping center, you will receive, afterwards, after the visit, a newsletter that is really the stuff that you actually interested in. So, through the whole customer journey. Are you a Swiss company? Yes, we are a Swiss startup. We just started 12 months ago. We grew from four to 28 people now already. 28? Yes, yes. So, some people really like this idea? Yes, definitely. I mean, we just started the business development two months ago and we are talking to a lot of big companies, banks, of course, also advertising in the street, but we are a software company. We want to develop the software, improve it, and then it can be used by many other companies. Right now, here, it looks like an Xbox system here. Maybe you should develop a specific camera. We have, we just have the Xbox camera here for the showcase. You have your own hardware? Yeah, we are not producing hardware, but we are on an early stage, still, as a startup. We have already the customers, but what we are trying to now to play with different hardware, because in the end, it should be a plug-and-play solution. So, we want that it's going to be possible to connect any camera. It's important to have a 4K hybrid solution so you get a lot of detail. You need a fast CPU arm. Yes, exactly. You need a lot of... Especially GPU. GPU. So, Nvidia is a good solution. Maybe the X1 or X2, the Jetson TX, you try these. Actually, we are not the techies in the company, but I guess we are up to date about every hardware that is available on the market. But what we also know to get back to your question, if you use 2D cameras, we need more power, more better GPU. So, we are working all together with Nvidia. It's good to have a 2D, two cameras. It helps to have two cameras, a 3D kind of system. Yeah, we are working right now on a system that uses different 2D cameras to make a 3D picture out of it. And you get the best result. You get the average of whatever this camera thinks and the other one, and then you'll get maybe more accurate age, maybe. Actually, this is not what we do. But the estimation is... How many... I think 10 estimations per second are done. Nice. Cool, so looking forward to better advertising everywhere, right? It's gonna be great.