EDGE AI POD
Discover the cutting-edge world of energy-efficient machine learning, edge AI, hardware accelerators, software algorithms, and real-world use cases with this podcast feed from all things in the world's largest EDGE AI community.
These are shows like EDGE AI Talks, EDGE AI Blueprints as well as EDGE AI FOUNDATION event talks on a range of research, product and business topics.
Join us to stay informed and inspired!
EDGE AI POD
Simple Cost Effective Vision AI Solutions at the edge
Sony's revolutionary IMX500 stands at the forefront of a quiet revolution in edge computing and smart city technology. This isn't just another image sensor—it's the first to integrate AI processing directly on the chip, transforming how visual data becomes actionable intelligence while preserving privacy and minimizing infrastructure requirements.
The power of this innovation lies in its elegant simplicity. Rather than sending complete images to cloud servers or external GPUs for processing, the IMX500 performs AI inference locally and transmits only the resulting metadata. This approach slashes bandwidth requirements to mere kilobytes, dramatically reduces power consumption, and—perhaps most critically—protects individual privacy by ensuring that identifiable images never leave the device. For urban environments where surveillance concerns often clash with safety imperatives, this represents a breakthrough compromise.
Real-world deployments already demonstrate the technology's transformative potential. In Lakewood, Colorado, where a one-mile stretch of road had become notorious for traffic fatalities, Sony's solution achieved 100% performance in identifying dangerous situations—outperforming three competing technologies while costing less. Through partnership with ITRON, these sensors can be seamlessly deployed using existing streetlight infrastructure, creating mesh networks of intelligent sensors without requiring expensive new installation work or dedicated power sources. This practical approach to deployment makes citywide implementation financially viable even for budget-constrained municipalities.
The implications extend far beyond traffic monitoring. From retail analytics to manufacturing quality control, the same core technology can be applied wherever visual intelligence provides value. By bringing AI to the edge in a form factor that addresses privacy, power, and practical deployment challenges, Sony has created a foundation for the next generation of smart infrastructure. Explore how this technology could transform your environment—whether an urban center, commercial space, or industrial facility—by leveraging the power of visual intelligence without the traditional limitations.
Learn more about the EDGE AI FOUNDATION - edgeaifoundation.org
So my presentation is going to focus on the sensor. As you may know, sony is leader in image sensor, but this new image sensor that we have is even different than what we had before. So the sensors that we had before it was just an image sensor. The new one that we have not only is an image sensor, it has AI integrated with it, so it makes it what Edge device, so you can do inferencing right on the device. I don't think anybody talk about silicon sensor having AI. So this is what Sony is bringing to Edge and is trying to bring some solution for smart applications. So let's talk about what is this sensor, this sensor called IMXY500. Imxy500, we work with different OEM companies to package this camera, so we are open. If you are an OEM company, you want to partner with us, please come, come talk to us. And then we have a platform called iTrios. Itrios is a platform that you can build your vision solution end-to-end. What does it mean? It means you have your camera, you take pictures, you build your data set, you have your data set, you need to train it. You train your data set and if you are happy with it, if the performance is good, you deploy it on the device After you deploy it on the device. That's not the end of the story, because we want to deploy like 10,000, 100,000 of devices. Then what do you need? Device managers, then you need to connect to the cloud. So the full path from the pictures all the way to deployment, scaling the solution. You can use our ITRIOS platform, but this is not just a platform, this is not just a sensor like any other edge devices.
Speaker 1:The significant features that you have for IMX 500 and this platform is what you see on this side. First of all is high speed, because it's at the edge, of course, right, we are all here because of the edge, right? So this is at the edge. And also is low bandwidth and low power. This is very interesting because I've heard a lot of people talking about low power. I want to have it like low bandwidth because it's expensive. How we achieve that? Because we don't send the image, we send the metadata. So look at the size that's shrinking to just, you know, a few kilobytes. So this is what is good about IMX 500. And the last thing is, when you send metadata, you don't send the full pictures. I have an application, I have people. They don't want the identity to be out there. So with the metadata that you send out there, you also protect their privacy. So this is the overview of IMX 500. Now you know the sensor, you know the platform.
Speaker 1:Let's talk about some of the applications that you can. Oh, my God, if I can figure this thing out Now. It's going fast, okay. Okay, sorry about that, I'm not Okay. So these are some applications that I put out there Today. I'm going to focus only on smart city. Okay, but this is a smart eye. You want to do an application for retail? Go for it. You want to do some application for logistic? Go for it. So this is like anywhere that you need to have a smart eye application, you can use IMX 500. But today we are going to talk about smart city, and smart city is kind of like now we see a lot of attention. We wanted to make our city more smarter and we are looking at some applications that we can deploy and, as you may know, this smart city application is not only one or two, it's the whole city. So then you need to be wise.
Speaker 1:How you install this, what is the cost of your solution? Let's talk about what is available today. If you look at the conventional system that is available today, you see the IP camera and it's bundled with a GPU, right. And then, right there, gpu is very expensive, right? So GPU, and then, when you have this, it has to go outside. When it goes outside, you need to protect it from rain, from sun, from snow and all of those things. So, as you see in the bottom, you have to put it in a box. You have to do some kind of structure to protect it. Look at that. That's cost, cost, cost, right.
Speaker 1:But now Sony is just giving a solution all in one. What does it mean? So, as you see here, we partner with Leopard and we built this camera. So this camera has two sensors and these two sensors you can adjust the angles when you install it. It's not a remote controller, but you can adjust the angle mood controller, but you can adjust the angle. And also, it's very easy to deploy, because in the city it's not easy to just go on top of a light and just install this. But this has brackets. You can just put it on the post, it can go vertical, it can go horizontal, so it's very easy to deploy.
Speaker 1:I told you it's low power, because this is metadata when it's low power, it means that I can get the power. It's not power hungry like a GPU, so that's another good thing. And also metadata, so we have privacy for protecting your identification. It's most of the city application you see people walking around right and also is very low cost. So, god, help me going to the next slide. Oh, perfect, I'm learning. Okay, so now the good news is that's not just a solution that I told you.
Speaker 1:Now, this solution has been released by Leopard, so okay, so this has been deployed, and this has been deployed in the city of San Jose. And also, this is easy to install. It has all of these good, significant features that I mentioned earlier. So Leopard is releasing this. This is available. But you are not only limited to applications that we have right now for a smart city. Okay, so you can. Now we have, like, traffic accounting or we have pedestrian or we have road safety, but this is a very open platform. You can deploy any model. I told you you can take pictures with model. If you have your own model, that's okay, bring it in, you can deploy it. You can build your solution. So any application can be deployed on that Leopard device that I told you, and now I have a very good example.
Speaker 1:This is actually concerning the safety. So in the city of Lakewood in Colorado, they were concerned because in the stretch of one mile of the road they had a lot of fatality, a lot of accidents. And then the city was concerned and they wanted to find a solution. They didn't care if the solutions are from Sony, it's from NVIDIA, it's from who. They didn't care, they just wanted to find a solution and avoid this fatality in that area. What they did? They compared four solutions. If you look at here, the first solution was LDAR and Edge application. Two companies participated in this and then this. So the third solution was camera and Edge application, and so this is another solution that they have, and the last one was Sony. Among all this, sony performed 100%, and you think, okay, 100%, how much do I have to pay for it? It was less expensive than all other solutions. So you see that it's not just a smart city, it's not me offering technology, I'm saving people's lives.
Speaker 1:So this is a very good story I wanted to share with you. I want to go to my last slide. Then I turn it over to my best friend here, vijay. So now, so far you learn about the camera, about all this great team's performance, but what I told you, remember? I said this is a very low power device. What does it mean? All the lights that you have outside, it has power If you can connect to the power and network from the light. What do you do? You make it less expensive. Deployment is going to be easier, installation is going to be easier and the magic of that is coming from iTron. So iTron has this possibility. They have an adapter that they can basically give us this possibility to deploy our cameras, connect them to the power and the network, and then we can build like mesh type of group of this network of these devices and deploy it.
Speaker 1:Please come. Thank you so much. Oh, let me mention this. So tomorrow we have another talk about Sony, but we are talking, we are focusing on manufacturing. If you are interested, it as 2015. Also, for the smart city, we have a full webinar happening with HAI Foundation on April 29. It's going to be announced on their page, so please join us If you have questions at all. We are in the corner, so I'm going to pass. Sorry if I took so much of your time. Okay, thank you so much, vijay. Do you think, testing, testing?
Speaker 2:Testing. Yeah, it's working. Good Good morning. And as we started to work with Sony a couple years ago on this solution and this basically will provide you how this can be deployed in real life here I see a sign connecting AI to real world, and so here's one of the examples we'll give you that how this gets connected, and we've been in this business from connecting devices, whether it's lighting controls or other kind of sensors and all that. But joining that with Sony on the AI side actually opens up a lot more opportunities than originally when we started this whole thing together. So my name is Vijay and I'm a product manager at. My name is Vijay and I'm a product manager at ITRON. I've been working on smart city smart lighting projects for many, many years, so adding this thing actually became much more interesting for me as well on the project side.
Speaker 2:So first is, what is the challenges that cities are seeing? Why do cities want any kind of sensors? What challenges they have? So, as you can know, the climate continues to change and I just have a list about four or five items that cities sees these as a challenges. I don't wanna read through the slides, every word, but you can. You'll get the presentation later. But the idea behind is that cities are seeing real problems, that they need to have sensors deployed in the cities. How, basically, can the smart city come through? So here's the challenges. And then what are the opportunities? How do we, as a technologist technology companies, how can we solve that? So? So, introducing the solution that we provide to the cities First is obviously there has to be devices and sensors.
Speaker 2:You start with that. That's where the data collection starts. Then intelligent connectivity Even if you have a sensor in one location, you gotta get the data somewhere. So intelligent connectivity is the next important feature Business intelligence why would customer pay for all these things? So they have to have understanding of, okay, what data I'm getting, how it's going to help me in overall solving the problem. And now, no one company can build all the products. We. We know that and that's been our motto from day one. We always bring the whole ecosystem of companies together and then go approach the customer to solve the challenges that earlier I talked about. So that's where our overall when we say ITRON, cityed, city edge is what that is taking all the solution and combining that.
Speaker 2:So we, we have been focusing on smart city applications for the last five, seven years and we basically have defined five major domains smart, smart lighting we've been doing that for almost 15 years now and a lot of cities have already deployed that. And then customers come back and say we want to add more beyond the lighting only, and that's where, basically, we start to add more and more sensors in the field with existing networks. So the next one is traffic and highways. This is where one of the partnerships with Sony has come together to make it much more AI-based solution available in the market. Public safety is also it will be part of the AI solution that's working with Sony, and then a few other ones I have. We have environmental sustainability, where the floods and air quality things come together, and a lot of our customers are utility customers, so they worry about the water metering and electricity and all the utility kind of applications. So this all comes into one platform and I can show you the architecture-wise how different devices in the field gets deployed. And then there's an ITRAN, intelligent Edge operating system where it takes the data and then go to the actual applications. That customer will be seeing what what customers need to try to solve, solve the challenges. It looks like a little messy picture, but you can see what.
Speaker 2:How the mesh networks work is that on the on the street light, you have the controller on the top and then you can add different sensors depending on what the needs of the customers are. It could be eye camera, it could be sensors, it could be flood sensors, weather stations a number of different applications can be put on and then, once it's connected, the data comes to the controller. Then it's a mesh network. Intelligently, it takes the data and sends that to the back office. Once the data is in the back office, then the intelligent applications can be built to provide the right outcome to the customers. So how does it get connected? You can see we have a little demo outside, and showing Simply is that Luminaire has the socket, what we call a seven pin, and then we're basically adding the adapter. The controller is already there. A lot of cities have already deployed that. You just add the adapter, which basically connects to the different sensors.
Speaker 2:Let's say this case is the AI camera. It provides the power to the camera as well as picks up the data. Once the data is picked up from there then goes to the mesh network and to come to the back software application where all this data can be metadata, as she talked about. Metadata is available and then can be used for the challenges that customers are seeing on whether there's a flooding, whether it's a traffic, whatever the problem is. So, in summary, adding the AI devices on the edge becomes on the smart city applications. It becomes much easier. Simplified installation you don't have to bring.
Speaker 2:One of the major problem customers see is that how do you provide the power to these sensors? So now we basically we're providing this network already on the street lights so we can take the power from there make the installation much, much simpler. Powering the device IoT devices some cases could be, you know, few watts of power, could be, you know, some milli watts of power, doesn't matter, all the powers you know makes it much, much simpler. You use the existing network from a communication point of view, so you don't really have to set up a new communication setup. Communications setup was already there and it's a scalable expand I was showing on the mesh network.
Speaker 2:This number of different devices can go on exactly the same network and the data can decipher at the back end which application, what problem is being solved and at the end is really the thing is low cost of ownership Is at the end. To deploy tens of thousands of these sensors in the field has to be low enough cost, because the budget for every city is very tight and providing the total lower cost of ownership makes it much more useful for the customer as well as for all of our citizens. And, like I said, we have a little demo area. You can come and see what applications we're working on and that's basically what my if there's any questions, you can come up also and we can answer. Thank you,