Martin Krantz of Smart Eye

00:00 Overview

JOHN: Today in The Garage, we’re live at CES 2025 with Smart Eye. Let’s go.
JOHN: Welcome to The Garage. I’m John Heinlein, Chief Marketing Officer with Sonatus. We’re recording live at CES 2025 right in the Sonatus booth. We’re pleased today to welcome our good friend and partner, Smart Eye. Our guest today is Martin Krantz, Founder and CEO of Smart Eye. Martin, welcome to The Garage.

MARTIN: Thank you.

00:33 Meet Martin Krantz

JOHN: So, we always like to begin by getting to know our guests. Tell us a little bit about yourself.

MARTIN: Yeah. I’m 53 years old, an engineer, engineering/physics background, father of four, and I’m a foodie.

JOHN: That’s great. And what did you do before the company, before Smart Eye?

MARTIN: I came from the university and started to work for a startup in Gothenburg, Sweden, where I live. That was a spin-off from the university. We made fiber optic pressure sensors. They were extremely accurate and extremely small. You can make fiber optic pressure sensors that were thinner than a hair, like half the size of a hair. Amazing technology. Very few customers. So, I learned a lot about what not to do when having a startup. So, it was a good lesson for me.

JOHN: Fantastic. And you say — we always like to learn fun facts about you — you said you’re a foodie. Tell us more about that.

MARTIN: Yeah. So, yeah, I like to cook. And I also, quite recently, or a few years ago, I started fermenting things. So, I’m doing my own, you know, sauerkraut and kombucha and kimchi and what-not.
JOHN: Well, I’m a huge foodie as well. We were chatting about this. I’m a huge foodie as well. We could talk for a long time about food, but, just for fun, I’ll give a different fun fact, which is I used to be a homebrewer. Most people don’t know that. People are surprised about that. But I learned to homebrew when I was living in Oregon for a little while, and it’s good fun. And also fermentation of a slightly different kind, but very good.

Martin: Yeah. Yeah.

01:13 About Smart Eye

JOHN: So then tell us about Smart Eye. Smart Eye is a great partner with us. We’ll get into that later, but tell us about your company.

MARTIN: Yeah. So, we started 25 years ago. Me and my father started it together. It was based on his dream, actually. He called me up one day in May in ’99. Sunday morning, he calls me up, and he said, I had this dream, and his wife — this was not part of the dream, this was actually happening — his wife, she had a frozen shoulder. So she had pain in her wrist, in her elbow, and in her shoulder from using a computer mouse. But his dream was that she could control a computer mouse by just looking at the screen and wherever she was looking, that mouse cursor went there. And then she blinked instead of clicking, and she got well in her arm. And he said, wasn’t this a cool dream? Yeah, I said, great dream. Great if it would work. Yeah, but you’re working with these things. Can’t you have a look at it to see if it’s doable? So, I started doing some calculations and, you know, I thought we don’t break any laws of physics here. I mean, it should be doable. So, we started the company 25 years ago. Now, pretty soon we started to gear towards automotive. Gothenburg` is, you know…

JOHN: Of course.

MARTIN: It’s the automotive capital of Sweden or whatever you call it. So, you have Volvo and Volvo trucks. Saab back in the day and so on.

JOHN: Sure.

MARTIN: So, very early we started talking to automotive customers and found out that this could be a great safety feature if it were to work in a car. And then it took some time to make it work in a car. But now, you know, we’re in more than 2 million cars worldwide. It’s ramping up really fast now. The use of driver monitoring. Eye tracking in cars, it’s called driver monitoring, and it’s saving a lot of lives. So, it’s a great time right now.

JOHN: Wonderful. I mean, such a focus on safety going on right now and recognizing that there’s so much distracted driving and so on. And so, I think the opportunity is really huge. What are some of the customers you’re working with?

MARTIN: We have 22 OEMs and these are the ones I’m allowed to talk about.

JOHN: Sure.

MARTIN: Here at the show — we got released from the NDA from Nissan and Mitsubishi — so we have a Nissan car here with Smart Eye in it that we showcase. And we have Volvo, Polestar, GM, Porsche, and Audi.

JOHN: And many others that you’re not disclosing. That’s great. Fantastic.

MARTIN: That’s what I can disclose.
JOHN: It’s impressive. And what’s your business model? Do you usually engage with Tier 1s, or OEMs, or a combination?

MARTIN: Yeah. The absolutely most common thing is that we have a Tier 1, which is usually a big global Tier 1.

JOHN: Sure.

MARTIN: They make the hardware and they integrate our software into a larger software stack.

05:08 Smart Eye Facial Analysis

JOHN: It’s impressive stuff. And I was just over at your booth a short time ago, seeing your demonstrations. And what’s impressive about it — I was just so blown away — is all the many things you can detect.

MARTIN: Yeah.

JOHN: You’re using AI in rich ways.

MARTIN: Yeah.

JOHN: You know, you started your story talking about just eye tracking for mouse. And that’s something that, you know, I’ve seen and some handicapped people use and that’s a great thing. And even carpal tunnel which I’ve suffered from myself as well. But you’re going much beyond that.

MARTIN: Yeah.

JOHN: You’re really detecting almost intent and mood and emotion. And to think that you could do that is mind blowing.
MARTIN: Yeah.

JOHN: So, tell us about some of those sort of ancillary things you do that are really powerful.

MARTIN: Yeah. So, the vision for our company is that, if you have someone that knows you really, really well, like a best friend or a family member or wife or whatever, and that person can tell so much just by looking at you, right? Looking at your expression, facial expression, you know, they know so much about you, and you should be able to teach a computer to do the same thing. So, that’s basically what we do. Like you said, it started with eye tracking only 25 years ago. Now we just analyze everything that has to do with the face of a driver. From it, we can infer drowsiness. We can infer intoxication, and we can infer distraction.

JOHN: Distraction, yeah.

MARTIN: Distraction, of course, and many other things.
JOHN: And anger as well. Frustration.

MARTIN: Absolutely. Anger. Frustration. We are sort of, scratching to be able to soon, really, truly, detect stress. I mean, we can see some signs which are stress-related already today, but the actual stress level inside of you that’s where we want to go. So, we want to be able to soon be able to extract vital signs, like pulse and breathing and stuff. So. now we have only talked about, really, the face of the driver, like a camera with a fairly narrow field of view, looking at the face of the driver. What’s also coming as a trend now in the industry is that you have cameras that look at the entirety of the inside of the car. And, there are many tragic accidents every year where people forget a child in the car in the summer and it’s hot and children die every year because of this. Not in the future.

JOHN: Pets as well.

MARTIN: Pets as well, even more common. Not in the future. I mean, with a system like Smart Eye, where you have cameras, you can also combine it with other sensors, such as interior radar. This will not happen in the future. I mean, the car is intelligent enough for that not to happen. And also, much more mundane, you can forget your purse. You can forget, you know, your laptop or whatever. And my favorite is, if you forget your phone, you will get the text message.
JOHN: That’s a joke if you’re not paying attention.

08:08 Sonatus and Smart Eye Collaboration

JOHN: You know, look, I was going to suggest we talk about this later, but actually, since you just mentioned that, I’d love to maybe take a moment to talk about the collaboration we’re doing with you, and then we can pull back and talk a bit more about the technology. So, we’ve done a fantastic collaboration with you here at the show where Sonatus is demonstrating our Automator AI product, which is an orchestration platform to allow OEMs and Tier 1s to create capabilities and develop things without coding very powerfully. And what’s interesting, as a brief sidebar on Automator, and then I’ll show you why it connects to this. What Automator does is it allows the OEMs to take inputs from many subsystems of the vehicle. So, not just one subsystem, but across the vehicle. Things are happening in the vehicle, location, actions, and so on like that. And then take actions outside, either controlling the vehicle, controlling vehicle functions, calling APIs, sending notifications, alerts, whatever. So, very rich inputs, very rich outputs. And so, the collaboration we’ve done with Smart Eye is to take your vision occupant monitoring system, integrate it into our demonstration vehicle, which is really exciting. 1970 Ford Bronco, which we restored and converted to EVs. So, it’s really a showpiece just to really gather people’s attention. But we put it in there and we put it in that car because it’s very obvious that it wasn’t a factory-integrated thing. So, for both of us, it’s obvious to our customers that we did that. It didn’t come that way. And here’s the fun part. You mentioned just a moment ago about the whole cabin.

MARTIN: Yeah.

JOHN: So, we’ve shown, as an example, how you could use this to create a teen driver mode. Where teens, and for example, teenage drivers. My daughter, you know, will be going towards driving relatively soon. You want to make sure that they’re following the rules of the road and following your agreement on how they should be driving as they’re learning. And one of them, by law in California, is they’re not allowed to have passengers for the first year of driving. How do you enforce that? You know, they take the keys, they go around the corner, their friend hops in and now they’re breaking the rules. Well, using Smart Eye, we’ve shown how we can detect passengers, which is, you know, baked into your system. And then, using the combination of our two technologies, cause an alert to the driver, reminding them that they’re breaking the rules, and send an alert to the parent so the parent knows that they’re breaking the rules.
MARTIN: Yeah.

JOHN: So, just a simple example and will increase safety for new drivers.

MARTIN: Great use case. I had no idea about this law, actually, that teenage drivers can’t have passengers for the first year.

JOHN: Only your parents. Of course, you want the parents to drive with them.

MARTIN: Of course.

JOHN: But of course, if you had a teen with another teen, they’re going to be talking instead of focusing on the road.
MARTIN: Exactly. And it makes perfect sense. And, I didn’t know about it. And of course, it’s a new use case I didn’t know about for Smart Eye.

JOHN: And then we’ve talked about how you can extend that into other things. You talked about other things left behind and so on like that. So, I think it’s just the very beginning of what we’ll do together. I think there’s so many great possibilities for our combined technology. So, first, I wanted to thank you for your great collaboration in this demonstration and people who have seen it, and perhaps we’ll put a short clip into this episode just so people can kind of get a sense of what that looks like. It’s really fun. It’s really fun stuff.
(Car notification) “Passengers are not permitted in Teen Driving Mode. Please have the passenger exit the vehicle before driving. A notification has been sent to your mom”

11:11 Smart Eye AI Technology

JOHN: So, this technology is all based on some really clever AI and some clever analysis. You want to talk a little bit about the technologies you’re using underlying this?

MARTIN: Yeah, sure. I mean, back in ‘99, we were dreaming about putting a whole image into a neural network. Wouldn’t it be cool? But then when the whole deep-learning revolution came, it became possible to do so. But we still have a little bit of legacy, computer vision in our stack. But, of course, like everyone else, we use the latest and greatest AI methods to enhance our stack. And, we have a huge amount of data of people who are driving in actual cars. A large part of the difficulty is to have, in a car… I mean, if you have a camera, like here in the studio, the lighting is controlled. It’s easy to do computer vision.

JOHN: Sure.

MARTIN: But in the car, it’s shaking and the light changes constantly and things like this. And behavior, I mean, people do the craziest thing while they drive. But we have so much data on it, and now, these days, we have also started to generate synthetic data, but the behavior in the synthetic data that we generate is not synthetic. We take behavior from actual drivers in actual drives, feed them into the synthetic data tool where we can change the appearance and the context of the drive. We can change the car. We can change the camera location. We can change the lens. We can change all of that. So basically…

JOHN: You’re multiplying your training set.
MARTIN: Exactly. We have developed now a tool to change that gives us almost an infinite amount of data which is super valuable when you need to harden your algorithms and validate them. You have to show to OEMs that this actually works in all conditions. And the bar, nowadays, is set super high. I mean, they require 99 point something in accuracy from your algorithm. But yeah, it’s high tech.
JOHN: Oh, I can attest to that. We deployed your technology into our vehicle. We deployed it and I asked the team, how’s it going? They said it works great.

MARTIN: Yeah.

JOHN: Immediately, it worked great.

MARTIN: Every time.

JOHN: And our vehicle you’ve never seen before. There’s no other vehicle like this in the world. And so, you couldn’t have pre calibrated for it and it was super reliable, super solid. We’re really impressed. And I’ll tell you another funny story about the demo and then we can talk more broadly. We were giving the demonstration the other day — this is a true story — we’re giving the demonstration the other day, and our demo presenter was on the side explaining the passenger scenario that I told you about. And at that time, I was simultaneously giving a tour to a CEO of another partner company who really wanted to get in the driver’s seat. And we’re not actually letting anyone get in, but I’m like, okay, okay, you can get in. So, I open the door, and he gets in the driver’s seat during the demo, and the alert didn’t go off because the sensitivity knows that wasn’t the passenger. And so, while he’s sitting in the driver’s seat, then the passenger got in and the right thing happened.

MARTIN: Yeah.

JOHN: It was rock solid. So fun. So great.

MARTIN: Yeah.

14:49 European Safety Regulation

JOHN: So, we were talking a little bit more, earlier about legislation in California. But I know the legislation, in Europe especially, is really evolving to really make some of these technologies mandatory. Can you talk about some of the evolution you’re seeing?

MARTIN: There’s an act for traffic safety called General Safety Regulation, or GSR for short, in Europe that is mandating that, in 2026, in July, all cars that are sold in the European Union — in the European market enlarged, I mean, also in Switzerland and Norway and Britain — these cars need to have a driver monitoring camera. It’s mandated by law. It’s becoming like seatbelts. Like airbags. They have to be there. The OEMs do not have a choice anymore. One and a half years from now. The law is four and a half years old. But now it’s…

JOHN: We’re really coming up on it.

MARTIN: We’re getting close. Yeah, we’re getting close. Of course, this is great, great news for Smart Eye.

JOHN: Sure.
MARTIN: I mean, there are 16 million cars sold in Europe every year. And we have a huge, like a large share of it is Smart Eye’s market share. So, things are looking really good. And we believe that, just like with the seatbelt and just like with the airbag, this law will originate in Europe, but it will also spread to the rest of the world. It makes perfect sense. It saves lives.

JOHN: It’s so crazy. You know, if you watch old movies, you see a car that has no seatbelts.
MARTIN: Yeah.

JOHN: Or you see a car, especially, that has no headrest, because think of whiplash.

MARTIN: Yeah.

JOHN: Would you get in a car with no headrest? Not a chance. And it’s almost, I think, in a few years, what do you mean there’s no monitoring? Because you don’t have to do anything wrong to be tired.

MARTIN: Yeah.

JOHN: You’re working hard and you’re trying to do the right thing and you’re tired. Obviously, you know, intoxication and other things, those are different. But, you know, it’s a preventative thing for the collective good.

MARTIN: Yeah.

JOHN: And obviously you’re doing things like privacy protection and things like that for people’s privacy. I think it’s obviously the right thing to do.
MARTIN: Yeah. Absolutely. Yeah. So, I mean, these days very few cars have accidents for mechanical reasons. And, if they do, you have seatbelts and airbags and that will save most of the lives that would have been lost otherwise. But, I mean, people still get distracted, especially since the advent of the smartphone.
JOHN: Yes.

MARTIN: People get very distracted. And as you said, people get tired. And intoxication: it’s a big problem. And it has actually been growing here in the US since COVID. So, there was a low in terms of intoxication deaths on the road, before COVID.

JOHN: Yes.

MARTIN: And it has increased ever since. And the explanation is probably twofold that people have been, you know, drinking more or taking more drugs and been more distressed since the pandemic. And also that the enforcement, the police controls, they are going down and…

JOHN: Police are stretched thinly.

MARTIN: Yeah. They are spread thin. And as that goes down, the drunk driving goes up. And there’s a lot of deaths here. I mean, I think it’s only in the U.S, 13,000 per year.

JOHN: Yeah. Well, and also pedestrian deaths are part of that as well. The number of pedestrian fatalities is really increasing.

MARTIN: Yeah.

JOHN: Certainly, in the U.S. I suspect worldwide. But I know in the U.S that’s the case.

18:39 The Possibilities of Automation

MARTIN: Yeah. And cars, they are getting more and more autonomous, and level two plus autonomy is very common these days. So, there are all sorts of strategies. It’s not part of Smart Eye’s domain, really, the strategies of what to do when, when intoxication has been detected. But you can think of all different things. Limp mode. You can think about the car will drive you home on level two level, it’s still better than having a drunk driver.

JOHN: Absolutely. We were talking about — we don’t engage in sort of just autonomous driving, I mean, there’s a whole crowded space and we don’t engage in that. We provide, maybe, infrastructure that might support that, but we’re not involved — but in this demonstration we talked about, we’ve had many people ask lots of interesting questions.

MARTIN: Yeah.

JOHN: For example, as I mentioned, our Automator product, we’re very excited that one of our recent achievements was our Automator product has a safety module now. This is brand new. Just in December, it was certified ISO 26262 ASIL D (delta), the high standard for safety certification. So, you can use it for drive actions not just stationary actions. So that really gives our OEM customers and Tier 1 customers confidence to integrate it in a functional safety / safe system. But a question that someone came up with is, okay, so now because if you get in a car, by definition the car’s not moving.

MARTIN: Yeah.

JOHN: So, people say, well, if the passenger gets in the car for your teen driver in our example before.

MARTIN: Yeah.

JOHN: Could you immobilize the car? And we said yes, we could. So, you could say, not only am I going to warn you, mister teen driver not following the rules, not only am I going to warn mom, but I’m going to not let you drive them. And perhaps, let’s say, lock the car out or whatever you decide to do, it’s totally up to the OEM’s control.

MARTIN: Yeah.

JOHN: But that kind of rich infrastructure could allow you to do a range of things to sort of address the issue because — I don’t know about you, and hopefully my daughter will never listen to this episode — but pretty much if I tell her to do something, she will automatically do what I’ve told her not to do.
MARTIN: Typical teenager.

JOHN: So, if you basically sort of have a time delay for her to rethink her bad decisions, then possibly it would give her an opportunity to make the better decision in 15 minutes or something, I’d guess.

MARTIN: Yeah.

JOHN: Or time for the mom to call and say, so, honey, I noticed I got this alert, can you tell me about what’s going on? So, it’s interesting possibilities.

MARTIN: Yeah. We have falling asleep. That often happens on a highway because it’s kind of boring to go on a highway. And, of course, it’s really difficult. I mean, you can’t stop the car on the highway.
JOHN: Of course not.

MARTIN: So, what to do there? It’s interesting.

JOHN: But there you can do alerts.

MARTIN: You can do alerts. You can also have, like, an ultra high final alert if you really detect actual sleeping.

21:38 Detecting Driving Patterns

JOHN: Sure. Do you do anything with — another question that people have talked about — I’m talking about this teen example, but another thing is the growing elderly population.

MARTIN: Yes.

JOHN: Where distraction or sleepiness could be an issue or even detecting other patterns in driving that could indicate some medical problem.

MARTIN: Yes.

JOHN: Are you exploring that kind of area?

MARTIN: Absolutely. Yeah. We say it’s like the next frontier for us. And, not surprisingly, we have a lot of requests from our Japanese customers for these use cases. And, of course, Japan has a growing elderly population. And many, many other places.

JOHN: As does the world with the baby boom coming through.

22:15 Differences in China

JOHN: You know, we’ve been talking about the U.S., we’ve been talking about Europe, but I think about the landscape across the world. What’s interesting in our business is that you see very different laws and different expectations worldwide. The China market is also, the Asian market, a little different. Tell us about some of the differences in China.

MARTIN: So, in China you have 10 or 15 OEMs or something like that. I don’t even know exactly.

JOHN: It’s incredible proliferation.

MARTIN: In the rest of the world, you usually develop a whole new car in three years, and China typically does it in one year. So that, of course, goes for all the subsystems. And then it means in practice, as a supplier, you’re going to deliver what you have. You’re not going to deliver something that you developed during the development of the car. You take what you have and deliver it. And also, production cycles are shorter there. Maybe 7 or 8 years here in the West, and, quite often, they’re shorter in China. So, I think of it as a little bit of a wild west in China. The automotive industry is so new and it has never, you know, been pulled together by a few players. So, let’s see what happens. I heard here at CES when I discussed this with some people that it seems like all these OEMs, they all have support from a local government. So, that makes it harder for them to merge.

JOHN: It’s true. The system there is a little different. They’re certainly forcing the innovation to happen more rapidly, both in terms of in their country and I think it is also forcing, you know, around the world, people to try to see how they could accelerate their design cycle.
MARTIN: Yeah.

JOHN: Also, the laws in China are different. They have different distraction laws in China as well, right?

MARTIN: Yeah. They have a few different. There’s a few regional differences. They have…

JOHN: Eating, I think you said.

MARTIN: Yeah, it’s called GBT in China. And they include eating, drinking, and smoking, and talking on the phone as sort of unwanted…

JOHN: Things that are not allowed or discouraged?
MARTIN: Or at least they are worthy of warning about. I don’t know if they are not allowed or not. I guess they have drive-thrus in China.

JOHN: Because eating in one’s car in the U.S is quite normal.

MARTIN: You couldn’t have it in the U.S., no.

JOHN: That’s right. But it is a source of distraction, honestly.

MARTIN: Yeah, it is.

JOHN: So, I think it’s quite sensible, really.

MARTIN: Yeah. Yeah. It’s sensible, but if you keep your eyes vigilantly on the road, I think you can still have a burger, or so.

JOHN: I might have had a burger or two while driving. I might have, but I’ve worked very hard to be careful when I’m doing it. But it’s a risk.
MARTIN: But we at Smart Eye, we’re going to a Shanghai auto show in April.

JOHN: Yeah.

MARTIN: So, I heard when people came back from Beijing last year, they were amazed about how things had developed during Covid. So, I’m eager to see it with my own eyes.

JOHN: It’s really impressive to see what the Chinese OEMs are achieving. Great innovation. Great pace. And, I think it’s healthy for the industry to see what’s possible and forcing people to rethink the status quo.

MARTIN: I’m a big believer in the free market forces. And I think competition is always good for everyone. So, all the boats rise with the rising tide.
JOHN: Martin, I’ve really enjoyed this conversation. We’ve touched on so many topics. I bet we could talk for another half an hour, but let me thank you for coming by, and we look forward to a fantastic, continued collaboration with you and your company. Well, thank you for joining us.

MARTIN: Thank you.
JOHN: If you like what you’re seeing from this episode, please like and subscribe to see more like it, both from CES and also from our home studio. We look forward to seeing you again in another episode of The Garage very soon.