 Hello, I'm Nelushi and I'm an analyst from ID Tech X. Today we're speaking to cluster imaging. So who are we speaking to here? My name is Semyon Nisenzar and I'm the founder of cluster imaging. We have computing depth, accurate depth using multiple cameras. And this, we develop two prototypes. One is for DSLR. Another is Mini, which we do video. And this will be used for automotive. It could be also used for factory automation and for security. Our target is automotive. And we are better in terms of resolutions than LiDARs and RADARs and much cheaper. OK, that's really excellent. Yes. I have a couple of questions about this product. So are you using a standard stereovision strategy to get your 3D images? We have much better than stereo. We have multiple cameras, stereo is in two cameras. Oh, I see, I see. OK, how many cameras do you have? This is number of cameras. Right, OK. So for our image, you can shot that. This is a depth map of we shot several days ago, which show distances. The color shows distances. And it's very accurate. And it's produce video, which supposed to put in the car. And it's either complement LiDARs and RADARs, or it's better. For medium distance on the street, we are up to 150 meters. And it have average, very accurate images. OK, so you're proposing that in an autonomous vehicle you can have your camera system and perhaps complementing one or two other technologies. Yes. You also mentioned that we could use something like this in an industrial vision situation. Yes. So what would we be using this for? Is this for high value components, and you want a 3D, a precise 3D image of it? Is that what you're using? Basically, we create precise 3D imaging. And it could be for inspection. It could be done in IRVR for robotic motion, because we are much more accurate at existing technology. Because we have multiple cameras and our advanced patented algorithms. So we can produce very accurately the distance. So if you take a robot arm and move it, you need to exactly where you put it. We are doing much better than anybody else. OK, so if we are using this for an inspection application. Is this something we could mount onto a production line? Or would this be more aimed at testing and prototyping at scale? Anywhere. Whenever you can use the distance, it's fine. Right, OK. It also can be used for security applications, the same as automotive. We need to put, this is a regular sensors. But when we are planning for next industrial prototype with automotive grade cameras, which can do day and night all the time. And for security, it will be very good. Because for security, important to get rid of false positive. Since we know the distance, big truck far away is not a problem. Or cat waving tail near the camera is also not a problem. Just go to right distance. So but our main target at this moment is for automotive, autonomous driving. And semi-autonomous. Autonomous will be long ago, but there's still assistance. Like for example, our car, our system can be used very well for parking. Now you have nice images of parking, but you better have also distances. Just to make sure. Definitely. Right. And on the side, the cars right now put multiple cameras on the side and front and back. And we can work with other guys in all of this and replace some of them or work together. For example, radar is very good for all weather kind of things. But it's very poor resolution. Yes. And we have very good resolution. So we can compliment. Yes, if it's fog and snow, maybe radar will not work. And us, work a little bit better than radar, but very bad condition, radar will work. For radar, radar is good for a very long distance. But for medium distance, we are much better. OK. So what about the actual image processing software? Is that something you provide yourself or are you working in partnership with a specialist software developer? We are doing our own software. And we are using standard parts. OK. Right now, we are using PC and Android devices. But we are planning to develop industrial prototype using NVIDIA. OK, excellent. So be part of the ecosystem and NVIDIA and can compliment all other components. Like you have a radar and then you put point cloud and you can combine their information with our information. How many cameras do you have? OK, we have a patent for multiple cameras, multi-resolution. And this is 21. This one, the fourth time, is 5 megapixel. And this is 2 megapixel. And it's combined. For DSLR, we use all of them. For video, we use only five. OK. And you're saying this and this just shows the range of applications you can have. And the type of cameras you can integrate into your system. Yeah, we use standard cameras. We can replace if needed, based on the customer. Yeah, certainly. So what about this setup right here? How does that differ? And how many cameras is this system using? It's the same number of cameras. But it has a big DSLR camera. So it's a full-size DSLR image with depth. OK, excellent. And this is whatever we have here. OK. You said that was just a small smartphone camera. Yeah, it's a cell phone camera. But we want to replace them with automotive-grade camera and make industrial prototype. So we are starting up and looking for investment to make industrial prototype for automotive. Are those FPGAs? It's all of FPGA, but it will be done differently. Right now, some chips will merge multiple images together and we will start designing the new version, which is cheaper and more compact. And it will be like this, size of a couple inches. OK, that's really excellent. Thank you so much for that interview. Really appreciate all that information you've shared with us. Really great product. Thank you. Is this better than Tesla? I don't know. I believe it's better. You think you're the best? Yes. All right.